Content delivery method, related device and medium

By selecting the target object based on the matching degree of content type and object labels in content push, the problem of insufficient exposure of new or old authors in new types of content is solved, and the content reach efficiency and network resource utilization are improved.

CN118939864AActive Publication Date: 2024-11-12TENCENT TECHNOLOGY (SHENZHEN) CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202310525322.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-10
Publication Date
2024-11-12
Estimated Expiration
2043-05-10

AI Technical Summary

Technical Problem

In the content push, the existing technology lacks friendliness to new types of content from new or old authors, resulting in the allocation of less exposure in the traditional mode, reducing the content reach efficiency and network resource utilization.

Method used

By determining the conditions based on the type of content to be delivered and the number of content published by the object, extracting and extending keywords, and selecting the target object for delivery based on the matching degree of the object label, the efficient reach of the content is achieved.

Benefits of technology

It improves the exposure and reach efficiency of new types of content of new or old authors, improves network resource utilization, and enhances the automation and accuracy of content delivery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939864A_ABST
    Figure CN118939864A_ABST
Patent Text Reader

Abstract

The invention provides a content delivery method, a related device and a medium, and belongs to the technical field of big data. The content delivery method comprises the following steps: determining that a first condition is established based on the number of published contents of the same type as to-be-delivered first contents and the number of published contents of a first object for delivering the first contents; in response to the establishment of the first condition, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword into a target keyword; obtaining an object label of each object in an object set on the content delivery platform; selecting a target object from the object set based on the matching degree of the target keyword and the object label; and putting the first content based on the selected target object. According to the embodiment of the invention, the object reaching efficiency of unpopular content delivery is improved, and the utilization rate of network resources is improved. The embodiment of the invention can be applied to various scenes such as information release, big data and artificial intelligence.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of big data, and in particular to a content delivery method, related devices and media. Background Art

[0002] The Internet involves a large amount of content push or delivery. At present, content push or delivery generally adopts the method of finding more popular content that meets the preferences of the target based on the target's browsing history or interest preferences, that is, "people looking for things". This is extremely unfriendly to some new content authors or new types of content by old authors. Since they lack the accumulation of historical delivery data, according to the traditional "people looking for things" method, very little exposure will be allocated to the content, which is not conducive to the efficiency of reaching the target object of specific content, so that some content has to reach the target object through additional channels, wasting network resources and reducing network resource utilization. Summary of the invention

[0003] The disclosed embodiments provide a content delivery method, related devices, and media, which can improve the efficiency of reaching targets for unpopular content delivery and improve network resource utilization.

[0004] According to one aspect of the present disclosure, a content delivery method is provided, comprising:

[0005] Determining that a first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content;

[0006] In response to the first condition being met, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword as a target keyword;

[0007] Obtain object labels for each object in an object set on a content delivery platform;

[0008] Selecting a target object from the object set based on a matching degree between the target keyword and the object tag;

[0009] The first content is delivered based on the selected target object.

[0010] According to one aspect of the present disclosure, there is provided a content delivery device, comprising:

[0011] A condition determination unit, configured to determine that a first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to which the first content is delivered;

[0012] a keyword determination unit, configured to, in response to the first condition being met, determine a first keyword based on the first content, expand a second keyword based on the first keyword, and integrate the first keyword and the second keyword as a target keyword;

[0013] A label acquisition unit, used to acquire an object label of each object in an object set on a content delivery platform;

[0014] an object selection unit, configured to select a target object from the object set based on a matching degree between the target keyword and the object tag;

[0015] A content delivery unit is configured to deliver the first content based on the selected target object.

[0016] Optionally, the condition determination unit is specifically used to:

[0017] If the number of published contents of the first object that delivers the first content is less than a first threshold, determining that the first condition is met;

[0018] determining a first type of the first content;

[0019] Determine the number of published contents of the first type on the content delivery platform;

[0020] If the number of published contents of the first type is less than a second threshold, it is determined that the first condition is met.

[0021] Optionally, the keyword determination unit is specifically used to:

[0022] Extracting a first sub-keyword from the text description of the first content;

[0023] extracting second sub-keywords from a text description of a second content placed by the first object before the first content;

[0024] Extracting third sub-keywords based on the multimodal analysis of the first content;

[0025] The first sub-keyword, the second sub-keyword, and the third sub-keyword are integrated into the first keyword.

[0026] Optionally, the keyword determination unit is specifically used to:

[0027] Inputting the first keyword into a related word prediction model to obtain a fourth sub-keyword;

[0028] The first keyword is expanded using the knowledge graph to obtain the fifth sub-keyword;

[0029] The fourth sub-keyword and the fifth sub-keyword are integrated into the second keyword.

[0030] Optionally, the target keyword has a first weight, the object tag has a second weight, and the object selection unit includes:

[0031] a matching degree determining unit, configured to determine a matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag and the second weight;

[0032] A target object selection unit is used to select a target object from the object set based on the matching degree of each object in the object set.

[0033] Optionally, the second weight includes a first sub-weight from a first source and a second sub-weight from a second source, and the matching degree determination unit is specifically configured to:

[0034] Determining a first matching degree corresponding to the first source based on the target keyword, the first weight, the object tag, and the first sub-weight;

[0035] determining a second matching degree corresponding to the second source based on the target keyword, the first weight, the object tag, and the second sub-weight;

[0036] The degree of matching is determined based on the first degree of matching and the second degree of matching.

[0037] Optionally, the first weight of the first keyword is a fixed weight, and the content delivery device further includes a weight determination unit, which is specifically configured to:

[0038] Determine the first weight of the second keyword, wherein the first weight of the fourth sub-keyword is the confidence probability of the fourth sub-keyword output by the associated word prediction model; and the first weight of the fifth sub-keyword is determined based on the distance between the fifth sub-keyword and the first keyword in the knowledge graph.

[0039] Optionally, the matching degree determination unit is specifically used to:

[0040] For each of the objects in the object set, determining a target object tag in the object tags of the object that matches one of the target keywords;

[0041] For each target object tag, determine a product of the second weight of the target object tag and the first weight of the matching target keyword;

[0042] The products of each target object label are accumulated to obtain the matching degree.

[0043] Optionally, the target object selection unit is specifically used to:

[0044] In the object set, determining candidate objects whose matching degree is greater than a third threshold;

[0045] Sorting the candidate objects in descending order of the matching degree;

[0046] The candidate object in the first position in the ranking is used as the target object.

[0047] Optionally, the label acquisition unit is specifically used to:

[0048] For each of the objects in the object set, obtaining a content set of interest represented by the object;

[0049] Obtaining a third keyword of the content in the content set;

[0050] For each of the third keywords, determining the number of contents in the content set having the third keyword;

[0051] If the number of contents is greater than a fourth threshold, the third keyword is used as the object tag of the object.

[0052] Optionally, the object selection unit includes:

[0053] A first screening unit, configured to perform a first screening on the object set based on a matching degree between the target keyword and the object tag, to obtain a first screened object;

[0054] A screening condition acquisition unit, used to acquire a first screening condition of the first object;

[0055] The second screening unit is used to perform a second screening on the first screened objects based on a match between the first screening condition and the attribute information of the first screened objects to obtain the target object.

[0056] Optionally, the second screening unit is specifically used for:

[0057] Based on the matching of the first screening condition and the attribute information of the first screened object, performing a second screening on the first screened object to obtain a second screened object;

[0058] Obtaining attribute information of the second filtered object;

[0059] Displaying a distribution diagram of the second filtered objects based on the attribute information;

[0060] Displaying an attribute information checkbox to receive a second filter condition input by the first object based on the distribution graph;

[0061] Based on the matching of the second screening condition and the attribute information of the second screened object, a third screening is performed on the second screened object to obtain the target object.

[0062] Optionally, the second screening unit is specifically used for:

[0063] Based on the matching of the screening condition and the attribute information of the first screened object, performing a second screening in the first screened object to obtain a second screened object;

[0064] sorting the second screened objects from high to low according to the matching degree;

[0065] taking the second screened object in the top second position in the ranking as the first display object, and displaying the first display object and the object label of the first display object matching the target keyword;

[0066] Selecting some of the second screened objects from the second screened objects that are not ranked in the second top position in the ranking as second display objects, and displaying the second display objects and the object tags of the second display objects that match the target keyword;

[0067] receiving a reserved number input by the first object, wherein the reserved number is selected by the first object according to the first display object, the object tag of the first display object matching the target keyword, the second display object, and the object tag of the second display object matching the target keyword;

[0068] The second screened objects are retained according to the retained number to obtain the target objects.

[0069] Optionally, the content delivery device further includes a first updating unit, and the first updating unit is used to:

[0070] Acquire the interest expression of the target object after delivering the first content;

[0071] According to the interest expression of the target object, the first weight of the target keyword of the first content is updated, wherein the first weight is a ratio of a first number to a second number, the first number is the number of times the target object expresses interest in the target keyword after the first content with the target keyword is delivered to the target object, and the second number is the number of times the first content with the target keyword is delivered to the target object.

[0072] Optionally, the content delivery device further includes a second updating unit, and the second updating unit is used to:

[0073] Acquire the interest expression of the target object after delivering the first content;

[0074] According to the interest expression of the target object, a second weight of the object tag of the target object is updated, wherein the second weight is a ratio of a third number to a fourth number, the third number is the number of times the target object expresses interest after the first content is delivered to the target object with the object tag, with respect to the object tag, and the fourth number is the number of times the first content is delivered to the target object with the object tag, with respect to the object tag.

[0075] Optionally, the content delivery device further includes a display unit, and the display unit is used to:

[0076] Acquire the interest expression of the target object after delivering the first content;

[0077] Acquire attribute information of the target object;

[0078] For each piece of the attribute information, determining a ratio of the target object corresponding to the attribute information that indicates interest;

[0079] The ratios corresponding to the respective attribute information are displayed for screening the delivery of the third content of the first object after the delivery of the first content.

[0080] According to one aspect of the present disclosure, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the content delivery method as described above when executing the computer program.

[0081] According to one aspect of the present disclosure, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the content delivery method as described above is implemented.

[0082] According to one aspect of the present disclosure, a computer program product is provided. The computer program product includes a computer program. The computer program is read and executed by a processor of a computer device, so that the computer device executes the content delivery method as described above.

[0083] In the disclosed embodiment, when the first condition is determined to be satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content, it means that the author of the first content is relatively new, or the type of the first content is relatively new. At this time, due to the lack of accumulation of historical delivery data, the first content will be allocated less exposure using the traditional mode, which reduces the efficiency of content reach. The disclosed embodiment determines the first keyword based on the first content, and expands the second keyword, which is integrated into the target keyword on this basis, and selects the matching target object for delivery for the first content based on the matching degree of the target keyword and the object label of each object in the platform object set. This delivery method does not rely on the accumulation of historical data of the first object, and can quickly reach the target object that may be of interest for the first content, thereby improving the object reach efficiency of specific content, without having to reach through other additional channels, and improving the utilization rate of network resources. The method of automatically generating keywords improves the degree of automation of content delivery, and jointly determines the target object of delivery based on the basic first keyword and the expanded second keyword, thereby improving the accuracy of content delivery.

[0084] Other features and advantages of the present disclosure will be described in the following description, and partly become apparent from the description, or be understood by practicing the present disclosure. The purpose and other advantages of the present disclosure can be realized and obtained by the structures particularly pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] The accompanying drawings are used to provide further understanding of the technical solution of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the technical solution of the present disclosure and do not constitute a limitation on the technical solution of the present disclosure.

[0086] Figure 1 is a schematic diagram of the system architecture of a system to which the content delivery method according to an embodiment of the present disclosure is applied;

[0087] Figure 2A -K is a schematic diagram of the content delivery method according to an embodiment of the present disclosure applied in a video delivery scenario;

[0088] Figure 3 is a flow chart of a content delivery method according to an embodiment of the present disclosure;

[0089] Figure 4 Shows Figure 3 Flowchart in which step 310 determines that the first condition is met;

[0090] Figure 5 Shows Figure 3 Flowchart of step 320 determining a first keyword based on the first content;

[0091] Figure 6 Shows Figure 5 A schematic diagram of an implementation process of extracting third sub-keywords based on multimodal analysis of the first content;

[0092] Figure 7 Shows Figure 3 Step 320 is a flowchart of expanding a second keyword based on the first keyword;

[0093] Fig. 8A -B shows Figure 7 A schematic diagram of an implementation process of expanding a second keyword based on a first keyword;

[0094] Fig. 9 Shows Figure 3 A schematic diagram of a keyword list formed by integrating target keywords in step 320;

[0095] Fig.10 Shown in Figure 3 Flow chart of adjusting target keywords after step 320;

[0096] Fig.11 Shows Figure 3 Flowchart of obtaining object labels of each object in step 330;

[0097] Fig.12 Shows Figure 3 A schematic diagram of an object tag list formed by integrating object tags in step 330;

[0098] Fig.13 Shows Figure 3 Step 340 is a flowchart of selecting a target object from an object set based on the matching degree between the target keyword and the object label;

[0099] Fig.14 Shows Fig.13 Flowchart of determining the matching degree between the target keyword and the object tag in step 1310;

[0100] Fig.15 Shows Fig.13 Step 1320 is a flowchart of selecting a target object from a set of objects;

[0101] Fig.16A -C shows Fig.13 A schematic diagram of an implementation process of selecting a target object from an object set in step 1320;

[0102] Fig.17 Shows Figure 3 Another flow chart of selecting a target object from the object set based on the matching degree between the target keyword and the object label in step 340;

[0103] Fig.18 Shows Fig.17 Step 1730 is a flowchart of performing a second screening on the first screened objects to obtain target objects;

[0104] Fig.19A -B shows Fig.17 A schematic diagram of an implementation process of performing a second screening on the first screened objects to obtain a target object in step 1730;

[0105] Fig. 20 Shows Fig.17 Another flow chart of step 1730 in which the first screened objects are subjected to a second screening to obtain target objects;

[0106] Fig.21A -C shows Fig.17 Another schematic diagram of the implementation process of step 1730 of performing a second screening on the first screened objects to obtain target objects;

[0107] Fig. 22 is a flow chart of a content delivery method according to another embodiment of the present disclosure;

[0108] Fig.23 is a flow chart of a content delivery method according to another embodiment of the present disclosure;

[0109] Fig.24 is a flow chart of a content delivery method according to another embodiment of the present disclosure;

[0110] Fig.25 Shows Fig.24 A schematic diagram showing the implementation process of the ratio corresponding to each attribute information is shown in FIG.

[0111] Fig.26 is a schematic diagram of a specific implementation process of a content delivery method according to an embodiment of the present disclosure;

[0112] Fig. 27 is a schematic diagram of the interaction process between an object and an object terminal of a content delivery method according to an embodiment of the present disclosure;

[0113] Fig.28 is a module diagram of a content delivery device according to an embodiment of the present disclosure;

[0114] Fig.29 is a terminal structure diagram of a content delivery method according to an embodiment of the present disclosure;

[0115] Fig.30 It is a server structure diagram of a content delivery method according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0116] In order to make the purpose, technical solution and advantages of the present disclosure more clear, the present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present disclosure and are not used to limit the present disclosure.

[0117] Before further describing the embodiments of the present disclosure in detail, the nouns and terms involved in the embodiments of the present disclosure are described. The nouns and terms involved in the embodiments of the present disclosure are subject to the following interpretations:

[0118] Artificial intelligence: It is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain target results. In other words, artificial intelligence is a comprehensive technology in computer science. It attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that machines have the functions of perception, reasoning and decision-making. Artificial intelligence technology is a comprehensive discipline that covers a wide range of fields, including both hardware-level technology and software-level technology. The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning. With the research and advancement of artificial intelligence technology, artificial intelligence technology has been studied and applied in many fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned driving, automatic driving, drones, robots, smart medical care, smart customer service, etc. I believe that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role.

[0119] Knowledge Graph: It is a modern theory that combines the theories and methods of applied mathematics, graphics, information visualization technology, information science and other disciplines with methods such as quantitative citation analysis and co-occurrence analysis, and uses visual graphs to vividly display the core structure, development history, frontier fields and overall knowledge architecture of the discipline to achieve the purpose of multidisciplinary integration. The main goal of the knowledge graph is to describe the various entities and concepts that exist in the real world, as well as the strong relationships between them. We use relationships to describe the association between two entities. From the perspective of artificial intelligence, the knowledge graph is a tool that uses knowledge bases to assist in understanding human language. From the perspective of databases, the knowledge graph is a method of storing knowledge in a graph. The knowledge graph is a relatively general formal description framework of semantic knowledge, which uses nodes to represent semantic symbols and edges to represent the relationship between semantics. The knowledge graph aims to describe various entities or concepts and their relationships in the real world. It constitutes a huge semantic network diagram, with nodes representing entities or concepts and edges consisting of attributes or relationships. The current knowledge graph has been used to refer to various large-scale knowledge bases. Knowledge graphs are also called semantic networks. Since the early days, semantic networks have promoted graph-based knowledge representation. For example, in the process of promoting the RDF standard, in such a graph-based knowledge representation system, entities are nodes of the graph, and the lines between nodes are relationships.

[0120] Entity: refers to something that is distinguishable and exists independently. For example, a person, a city, a plant, a commodity, etc. Everything in the world is made up of specific things, which refers to entities. Entities are the most basic elements in the knowledge graph, and different entities have different relationships.

[0121] Relationship: A certain relationship between entities. A relationship is formalized as a function that maps k points to a Boolean value. In the knowledge graph, a relationship is a function that maps k graph nodes (entities, semantic classes, attribute values) to Boolean values.

[0122] The Internet involves a large amount of content push or delivery. At present, content push or delivery generally adopts the method of finding more popular content that meets the preferences of the target based on the target's browsing history or interest preferences, that is, "people looking for things". This is extremely unfriendly to some new content authors or new types of content by old authors. Since they lack the accumulation of historical delivery data, according to the traditional "people looking for things" method, very little exposure will be allocated to the content, which is not conducive to the efficiency of reaching the target object of specific content, so that some content has to reach the target object through additional channels, wasting network resources and reducing network resource utilization.

[0123] System architecture and scenario description of the application of the embodiments of the present disclosure

[0124] Figure 1 It is a system architecture diagram used in the content delivery method according to the embodiment of the present disclosure, which includes an object terminal 140, the Internet 130, a gateway 120, a content delivery platform server 110, and an object tag library 150 of the content delivery platform server 110.

[0125] The target terminal 140 includes various forms such as desktop computers, laptop computers, PDAs (personal digital assistants), mobile phones, vehicle-mounted terminals, home theater terminals, and dedicated terminals. In addition, it can be a single device or a collection of multiple devices. The target terminal 140 can communicate with the Internet 130 in a wired or wireless manner to exchange data. The target terminal 140 includes a content delivery recommendation system 141.

[0126] When object A operates on the front-end interface of the object terminal 140, the object terminal 140 will input the content to be delivered uploaded by object A into the content delivery recommendation system 141, and the content delivery recommendation system 141 will match the content to be delivered with the object tags extracted from the object tag library 150 of the content delivery platform server 110, and select several objects from the content delivery platform server 110 that will deliver the content to be delivered according to the matching situation, and finally deliver the content to be delivered to the selected objects.

[0127] The content delivery platform server 110 refers to a computer system that can provide certain services to the target terminal 140. Compared with the ordinary target terminal 140, the content delivery platform server 110 has higher requirements in terms of stability, security, performance, etc. The content delivery platform server 110 can be a high-performance computer in the network platform, a cluster of multiple high-performance computers, a part of a high-performance computer (such as a virtual machine), a combination of parts of multiple high-performance computers (such as virtual machines), etc.

[0128] The gateway 120 is also called an internetwork connector or a protocol converter. The gateway realizes network interconnection at the transport layer and is a computer system or device that acts as a converter. The gateway is a translator between two systems that use different communication protocols, data formats or languages, or even have completely different architectures. At the same time, the gateway can also provide filtering and security functions. The message sent by the object terminal 140 to the content delivery platform server 110 must be sent to the corresponding content delivery platform server 110 through the gateway 120. The message sent by the content delivery platform server 110 to the object terminal 140 must also be sent to the corresponding object terminal 140 through the gateway 120.

[0129] The embodiments of the present disclosure can be applied in various scenarios, such as Figure 2A - Video delivery scenarios shown in K, etc.

[0130] like Figure 2A As shown, object A opens a social application on the object terminal 140 and enters the interface corresponding to the personal center in the social application. Among them, the interface corresponding to the personal center displays information such as browsing settings and my video number. My video number contains the nickname, avatar, number of followers, video number messages and other content of object A. The "Publish Video" button and the "Start Live Broadcast" button are displayed at the bottom of the interface. When object A clicks the "Publish Video" button, the video publishing process is entered; when object A clicks the "Start Live Broadcast" button, the live broadcast process is entered. At this time, object A clicks the "Publish Video" button to enter the video publishing process.

[0131] like Figure 2B As shown in the figure, when subject A clicks the "Publish Video" button, the "Shoot" button, the "Select from Album" button and the "Cancel" button will be displayed on the interface. Among them, the "Shoot" button is used for the subject to shoot a video in real time; the "Select from Album" button is used for the subject to select an existing video from the album to publish. At this time, subject A clicks the "Select from Album" button to enter the video selection process.

[0132] like Figure 2C As shown, when object A clicks the "Select from Album" button, multiple pictures and videos contained in the album will be displayed on the interface. Object A selects the first content in the album and clicks the "Next" button to confirm the selection of the first content for publication.

[0133] like Figure 2D As shown, when object A selects the first content and clicks the "Next" button, the interface will display the first content, an input box for adding a text description, and selection controls such as "Location", "Activity" and "Link". At this time, object A enters the text description of the first content in the input box for adding a text description and clicks the "Package" button to enter the delivery object screening process.

[0134] like Figure 2EAs shown, when object A clicks the "Dig Package" button, the delivery object screening process is entered. In the delivery object screening process, the content delivery recommendation system 141 in the object terminal 140 will receive the first content, extract keywords from the first content, and display the keyword list on the interface. The displayed keyword list contains six target keywords, which are keyword A, keyword B, keyword C, keyword D, keyword E, and keyword F. Among them, the weight of keyword A is 0.05, the weight of keyword B is 0.15, the weight of keyword C is 0.45, the weight of keyword D is 0.1, the weight of keyword E is 0.2, and the weight of keyword F is 0.05. Object A can click the "Modify" button to adjust the target keywords and weights in the keyword list, or click the "OK" button to go directly to the next step. At this time, object A can click the "Modify" button to adjust the target keywords and weights in the keyword list.

[0135] like Figure 2F As shown, when subject A clicks the "Modify" button, subject A can adjust the target keywords and the weights of each target keyword on the interface. First, subject A adjusts the target keywords in the keyword list to keyword Q, keyword B, keyword C, keyword D, keyword E, and keyword K. Next, subject A adjusts the weight of keyword Q to 0.2, the weight of keyword B to 0.05, the weight of keyword C to 0.4, the weight of keyword D to 0.05, and the weight of keyword K to 0.1. After the adjustment is completed, subject A clicks the "OK" button to enter the crowd screening stage.

[0136] like Figure 2G As shown, in the crowd screening stage, the interface of the object terminal 140 will display various conditional screening items of the conditions for the delivery population and the input boxes corresponding to the conditional screening items, wherein the conditional screening items include the activity level of the population, the education level and the population region. Object A enters "≥ medium activity level" in the input box corresponding to "Crowd screening", enters "≥ junior high school" in the input box corresponding to "education level", and enters "City A" in the input box corresponding to "Crowd region". At this time, the delivery population is the population with junior high school education or above in City A, and whose activity level is not lower than medium activity level. Object A clicks the "OK" button to confirm the screening conditions for the delivery population.

[0137] like Figure 2HAs shown, when object A clicks the "OK" button, the content delivery recommendation system 141 of the object terminal 140 will screen each object in the content delivery platform server 110 according to the received delivery population screening conditions, obtain the delivery population that meets the conditions, and generate an overall distribution map according to the attribute information such as the activity level and education level of the target delivery population. Then, the distribution of the delivery population that meets the conditions is displayed to object A. Specifically, there are a total of 80 delivery populations that meet the conditions. The overall distribution includes the activity ratio and the education level ratio. In the activity ratio, there are 50 people with high activity and 30 people with medium activity. In the education level ratio, there are 10 people with junior high school education, 40 people with college education, and 30 people with high school education. At this time, object A can click the "Previous Step" button to return to the setting link of the screening conditions and reset the screening conditions; object A can also click the "OK" button to determine that the 80 people who meet the conditions will be the final delivery population.

[0138] like Fig.2I As shown, when object A clicks the "OK" button, the content publishing process is entered. In the content publishing process, the interface will display the first content, the text description of the first content entered by object A, and selection controls such as "location", "activity" and "link". At this time, object A clicks the "Publish" button to confirm the publication of the first content.

[0139] like Figure 2J As shown, when object A clicks the "Publish" button, a prompt pop-up window will be displayed on the interface, and the prompt pop-up window displays the prompt field "Do you want to deliver the first content to the 60 objects that meet the conditions?" At this time, object A clicks the "Yes" button, and the object terminal 140 delivers the first content in response to object A's selection.

[0140] like Figure 2K As shown, after the delivery of the first content is completed, a prompt pop-up window will be displayed on the interface, and the prompt pop-up window will display the prompt field "Content delivery has been completed". At this time, object A clicks the "OK" button to confirm the completion of the publication of the first content.

[0141] General description of the disclosed embodiments

[0142] According to an embodiment of the present disclosure, a content delivery method is provided.

[0143] This content delivery method can be used to Figure 2A- The video delivery scenario shown in K can also be used for audio work recommendation, information recommendation and other scenarios. At present, the disclosed embodiment provides a solution for determining target keywords based on the content to be delivered, and selecting matching target objects for delivery based on the matching degree between the target keywords and the object labels of each object in the platform object set, which can quickly reach the target objects that may be interested in the content to be delivered, thereby improving the object reaching efficiency of specific content and the accuracy of content delivery.

[0144] like Figure 3 As shown, a content delivery method according to an embodiment of the present disclosure may include:

[0145] Step 310: Determine that a first condition is met based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content;

[0146] Step 320: In response to the first condition being met, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword as a target keyword;

[0147] Step 330: Obtain object tags for each object in the object set on the content delivery platform;

[0148] Step 340: Select a target object from the object set based on the matching degree between the target keyword and the object tag;

[0149] Step 350: delivering the first content based on the selected target object.

[0150] Steps 310 - 350 are briefly described below.

[0151] In step 310, the first object refers to an object that wants to share some content with other objects via the Internet. The first content refers to the video, picture, text, etc. that the first content wants to publish.

[0152] The number of published content refers to the total number of videos, pictures, and texts that the first object has published on the content delivery platform.

[0153] The first condition refers to the requirements that should be followed when delivering content based on keyword matching.

[0154] In step 320, the first keyword is a core information word that can describe the first content. The second keyword is a word that has a high degree of semantic similarity to the first keyword.

[0155] In step 330, the content delivery platform refers to a platform that enables each object to share the content that he wants to publish with other objects. For example, commonly used content delivery platforms include video accounts, public accounts, social media applications, etc.

[0156] Object label refers to the standardized definition of object attributes, behaviors, preferences, etc., which is used to describe the object characteristics of each object on the content delivery platform.

[0157] In step 340 , the matching degree is a feature that can describe the similarity between the target keyword and the object tag. The matching degree can be expressed in a numerical value, a vector or other forms.

[0158] The target object is an object on the content delivery platform that may be interested in the first content.

[0159] In step 350, the process of delivering the first content based on the selected target object includes but is not limited to the following steps:

[0160] Establishing a communication connection with the content delivery platform;

[0161] The first content delivered by the first object is delivered to the target object of the content delivery platform.

[0162] In the specific implementation of this embodiment, the target terminal first establishes a communication connection with the content delivery platform, and then delivers the first content to the target object on the content delivery platform according to the preset frequency and preset time, so that the first content is exposed to the target object, so that the target object can browse, click, download, forward or like the first content. Among them, the preset frequency and preset time can be set according to actual needs without limitation.

[0163] Furthermore, when delivering the first content to the target object, the first content may be displayed to the target object in the form of a pop-up window prompt, or the first content may be displayed in the center of the browsing page so that the target object pays attention to the first content first. The first content may also be displayed in other ways without limitation.

[0164] Through the above steps 310-350, in the embodiment of the present disclosure, when the first condition is determined to be met based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content, it means that the author of the first content is relatively new, or the type of the first content is relatively new. At this time, due to the lack of accumulation of historical delivery data, the first content will be allocated less exposure using the traditional mode, reducing the content reach efficiency. The embodiment of the present disclosure determines the first keyword based on the first content, and expands the second keyword, which is integrated into the target keyword on this basis, and selects the matching target object for delivery for the first content based on the matching degree of the target keyword and the object label of each object in the platform object set. This delivery method does not rely on the accumulation of historical data of the first object, and can quickly reach the target object that may be of interest for the first content, thereby improving the object reach efficiency of specific content, without having to reach through other additional channels, and improving the utilization rate of network resources. The method of automatically generating keywords improves the degree of automation of content delivery, and jointly determines the target object of delivery based on the basic first keyword and the expanded second keyword, thereby improving the accuracy of content delivery.

[0165] The above is a general description of steps 310 to 350. Since step 350 has been described in detail in the above general description, the specific implementation of steps 310, 320, 330 and 340 will be described in detail below.

[0166] Detailed description of step 310

[0167] In step 310, it is determined that the first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content.

[0168] In the specific implementation of this embodiment, when the first object is an object with a small number of publications, it indicates that the first object is a relatively new author. If the first content is directly delivered to the content delivery platform, the possibility that the first content published by the first object can reach the target object of interest is low. Therefore, it is possible to judge whether the first condition is met based on the number of contents published by the first object. When the first condition is met, the content delivery method of the embodiment of the present disclosure is adopted to increase the exposure of the content published by the first object, so that the content published by the first object can reach the target object that may be interested.

[0169] On the other hand, when the first object is published a large number of times, it indicates that the first object is an author with a certain degree of exposure. If the first object publishes a first content of a relatively new type, since there are fewer published contents of the same type as the first content, determining the target object of interest based on the type of the first content will affect the exposure of the first content. Therefore, it is possible to determine whether the first condition is met based on the number of published contents of the same type as the first content. When the first condition is met, the content delivery method of the embodiment of the present disclosure is adopted to increase the exposure of the first content of a relatively new type, so that the first content reaches the target objects that may be interested.

[0170] It should be noted that, in the embodiment of the present disclosure, when the number of published content of the same type as the first content to be delivered and the number of published content of the first object for delivering the first content are obtained, a separate permission or separate consent of the first object will be obtained through a pop-up window or by jumping to a confirmation page. After the separate permission or separate consent of the first object is clearly obtained, the number of published content of the same type as the first content to be delivered and the number of published content of the first object for delivering the first content are obtained.

[0171] like Figure 4 As shown, in some embodiments, the process of determining that the first condition is met in step 310 includes but is not limited to the following steps 410 to 440.

[0172] Step 410: If the number of published contents of the first object for delivering the first content is less than the first threshold, it is determined that the first condition is met;

[0173] Step 420: Determine a first type of the first content;

[0174] Step 430: Determine the number of published contents of the first type on the content delivery platform;

[0175] Step 440: If the number of published contents of the first type is less than the second threshold, it is determined that the first condition is met.

[0176] Steps 410 - 440 are described in detail below.

[0177] In step 410, a threshold comparison method can be used to determine whether the first object is a newer author. Specifically, the number of published contents of the first object that delivers the first content is obtained from the object terminal log, wherein the object terminal log is used to record the data generated by various object behaviors of the first object in the past, such as publishing behavior, browsing behavior, forwarding behavior, and liking behavior. Then, the number of published contents of the first object is compared with the first threshold to obtain a first comparison result. Finally, if the first comparison result is that the number of published contents of the first object is less than the first threshold, it indicates that the first object is a newer author and the exposure of the content published by the first object needs to be increased. Therefore, it is determined that the first condition is met. If the first comparison result is that the number of published contents of the first object is not less than the first threshold, it indicates that the first object is not a newer author, and the first object is an author with a certain amount of exposure. In order to make the first content published by the first object trigger the target object of interest as much as possible, it is necessary to further determine whether the first condition is met according to the type of the first content.

[0178] In step 420, the first object may select a type from the preset types displayed on the front-end interface of the object terminal as the first type of the first content, and the object terminal may determine the first type of the first content in response to the selection operation of the first object. Specifically, the preset type includes multiple levels of tags, and each first-level tag is also provided with a second-level tag. For example, the first-level tags include music, food, travel, film and television, sports, and the like. When the first-level tag is music, the subordinate second-level tags include pop music, rock music, folk music, and the like; when the first-level tag is food, the subordinate second-level tags include Chinese food, Western food, and the like. When the first object selects "Food-Chinese Food" from the multiple preset types displayed on the front-end interface of the object terminal, the object terminal may determine the first type of the first content as Chinese food.

[0179] In step 430, the target terminal may issue a data acquisition request to the content delivery platform to request the number of published contents of the first type on the content delivery platform. Specifically, the target terminal first sends a data acquisition request to the content delivery platform. The content delivery platform may query all content types belonging to the first type of published contents on the background server according to the data acquisition request, count the number of published contents of the first type, and send a response containing the number of published contents of the first type to the target terminal, so that the target terminal can determine the number of published contents of the first type on the content delivery platform.

[0180] It should be noted that, in the disclosed embodiment, when the object terminal sends a data acquisition request to the content delivery platform, it will obtain the separate permission or separate consent of the first object through a pop-up window or jump to a confirmation page, etc., and after clearly obtaining the separate permission or separate consent of the first object, it will send the data acquisition request to the content delivery platform. In addition, when the content delivery platform feeds back the number of published content of the first type according to the data acquisition request, it will also obtain the separate permission or separate consent of the first object through a pop-up window or jump to a confirmation page, etc., and after clearly obtaining the separate permission or separate consent of the first object, it will send the number of published content of the first type to the object terminal.

[0181] In step 440, after determining the number of published content of the first type, the object terminal compares the number of published content of the first type with the second threshold to obtain a second comparison result. If the second comparison result is that the number of published content of the first type is less than the second threshold, it indicates that the first type to which the first content belongs is a newer type and the exposure of the first content needs to be increased. Therefore, it is determined that the first condition is met. If the second comparison result is that the number of published content of the first type is not less than the second threshold, it indicates that the first type to which the first content belongs is a type that is well known to the objects on the content delivery platform, and a certain number of target objects interested in the first type have accumulated on the content delivery platform. The probability that the first content can reach the target objects of interest is high, and it is not necessary to use the content delivery method of the embodiment of the present disclosure to increase the exposure of the first content. Therefore, it is determined that the first condition is not met.

[0182] It should be noted that the first threshold and the second threshold mentioned above can be set according to actual conditions without specific limitation.

[0183] Through the above steps 410-420, the embodiment of the present disclosure can determine whether the first object delivering the first content is a newer author by using the number of published contents of the first object delivering the first content and a threshold comparison method, and when the first object is not a newer author, determine whether the first content is a newer type of published content by using the number of published contents of the first type on the content delivery platform and a threshold comparison method. This method can effectively determine the first content published by a newer author or a newer type of first content, which is conducive to quickly reaching the target objects that may be interested in the first content, thereby improving the object reaching efficiency of the first content.

[0184] In a specific embodiment, the process of determining that the first condition is met in step 310 includes but is not limited to the following steps:

[0185] Display trigger controls;

[0186] In response to the selection of the trigger control, it is determined that the first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object for delivering the first content.

[0187] The trigger control is a visual control on the front-end interface of the object terminal, and the trigger control is used to provide the first object with a selection. When the first object selects the trigger control, it will enter the determination process of whether the first condition is met.

[0188] In the specific implementation of this embodiment, after the first object triggers the content delivery process, the object terminal will display the trigger control to the first object on the interface. If the first object selects the trigger control, the object terminal will respond to the selection operation of the first object and determine whether the first condition is met based on the number of published contents of the same type as the first content and the number of published contents of the first object. When the first condition is met, the content delivery method of the embodiment of the present disclosure is used to deliver the first content.

[0189] The specific implementation process of this embodiment is similar to the above steps 410-420. The difference is that the determination process of the first condition in this embodiment is started after the first object triggers the content delivery process, based on the first object's selection of the trigger control, while the determination process of the first condition in the above steps 410-420 is directly started after the first object triggers the content delivery process. To save space, it will not be repeated.

[0190] This method can determine whether the first condition is met based on the first object's selection of the trigger control, so that the first object can independently choose whether to start the judgment process of the first condition, which can improve the personalization level of the content delivery method of the embodiment of the present disclosure.

[0191] Detailed description of step 320

[0192] In step 320 , in response to the first condition being met, a first keyword is determined based on the first content, a second keyword is expanded based on the first keyword, and the first keyword and the second keyword are integrated as a target keyword.

[0193] In the specific implementation of this embodiment, if the first condition is met, it indicates that the first content is published by a relatively new author or is a relatively new type of published content. If the first content is directly delivered to the object on the content delivery platform, the possibility of the first content reaching the target object of interest is low, which is not conducive to increasing the exposure of the first content. Based on this, the embodiment of the present disclosure considers a solution for determining the target object of the first content based on keyword matching, so as to improve the matching degree between the first content and the target object, thereby improving the accuracy of content delivery.

[0194] In actual application scenarios, when publishing content, the subject often enters a text description of the published content. These text descriptions are used to briefly introduce the published content and will cover the more critical and important content information of the published content. Therefore, the embodiment of the present disclosure considers a solution for automatically generating keywords for the published content based on the text description of the published content, which can improve the degree of automation of keyword extraction, thereby improving the convenience and accuracy of keyword extraction. In addition, the embodiment of the present disclosure also considers a solution for setting a corresponding weight for each keyword, which can improve the computational efficiency of the keyword matching process, thereby improving the efficiency of content delivery.

[0195] Combine the following Figure 5 and Figure 6 The process of determining the first keyword based on the first content is described in detail.

[0196] In some embodiments, the process of determining the first keyword based on the first content includes but is not limited to the following steps 510 to 540 .

[0197] Step 510: extracting a first sub-keyword from the text description of the first content;

[0198] Step 520: extracting second sub-keywords from the text description of the second content placed by the first object before the first content;

[0199] Step 530: extracting third sub-keywords based on the multimodal analysis of the first content;

[0200] Step 540: Integrate the first sub-keyword, the second sub-keyword, and the third sub-keyword into the first keyword.

[0201] Steps 510 - 540 are described in detail below.

[0202] In step 510, firstly, a text description of the first content input by the first subject on the interface of the subject terminal is obtained. Then, the first sub-keyword can be extracted from the text description of the first content by using machine learning. The process of extracting the first sub-keyword from the text description of the first content by using machine learning includes but is not limited to the following steps:

[0203] Perform word segmentation and part-of-speech tagging on the text description of the first content to obtain candidate keywords;

[0204] Use the pre-trained BERT module to semantically encode the candidate keywords and generate the candidate keyword vector corresponding to each candidate keyword;

[0205] For each candidate keyword, based on the candidate keyword vector, calculating the relevance between the candidate keyword and the first type of the first content;

[0206] The candidate keywords are sorted according to the relevance, and the candidate keyword ranked in the front predetermined position is selected as the first sub-keyword.

[0207] In the specific implementation of this embodiment, firstly, the text description of the first content is segmented according to conventional grammatical specifications, and the text description of the first content is split into segmentation data such as characters, words or phrases. Then, each character, word or phrase is marked with a part of speech. Then, the meaningless characters in the segmentation data are deleted, wherein the meaningless characters refer to the punctuation marks generated when the text description is segmented; according to the stop word list provided by the content delivery platform, the stop words in the segmentation data are deleted; according to the part of speech of the segmentation data, only nouns, adjectives and verbs with a word length of not less than 2 in the segmentation data are retained, and the retained nouns, adjectives and verbs with a word length of not less than 2 are used as candidate keywords.

[0208] Furthermore, the pre-trained BERT module is used to semantically encode the candidate keywords, extract the semantic information of each candidate keyword, and obtain the candidate keyword vector corresponding to each candidate keyword.

[0209] It should be noted that the pre-trained BERT module (Bidirectional Encoder Representations from Transformers, BERT) is a commonly used module in machine learning. The pre-trained BERT module of the embodiment of the present disclosure is implemented based on transformer, which is a bidirectional encoding model. When the pre-trained BERT module is used to semantically encode the candidate keywords, the candidate keywords are first forward encoded in a left-to-right order to obtain a forward encoding feature vector; then, the candidate keywords are reversely encoded in a right-to-left order to obtain a reverse encoding feature vector; finally, the forward encoding feature vector and the reverse encoding feature vector are vector fused to obtain the candidate keyword vector corresponding to the candidate keyword pair.

[0210] Furthermore, a preset Softmax function can be used to calculate the probability distribution vector of each candidate keyword vector on the first type, and the probability distribution vector is used as the relevance between the candidate keyword and the first type of the first content. The larger the probability distribution vector, the higher the relevance of the candidate keyword to the first type.

[0211] It should be noted that the Softmax function is a normalized exponential function that can "compress" a K-dimensional vector z containing any real number into another K-dimensional real vector σ(z) so that the range of each element is between (0,1) and the sum of all elements is 1. This function is often used in multi-classification problems.

[0212] Finally, the candidate keywords are arranged in descending order according to the order of relevance, so that the candidate keywords with higher relevance are ranked at the front and the candidate keywords with lower relevance are ranked at the back; according to the arrangement order, the candidate keywords ranked in the front predetermined ranking are selected as the first sub-keywords. Among them, the predetermined ranking is set according to the actual situation and is not limited. For example, the predetermined ranking can be 5, and the candidate keywords ranked in the top five in relevance are selected as the first sub-keywords.

[0213] In step 520, since the first object often has a certain amount of accumulated content in the past, the second content placed by the first object before the first content can also reflect the content placement preference of the first object to a certain extent. Therefore, the second content placed by the first object before the first content can also be used as a basis for determining keywords. Based on this, first obtain the text description of the second content placed by the first object before the first content from the object terminal log. Then, extract the second sub-keyword based on the text description of the second content.

[0214] Specifically, the specific process of extracting the second sub-keywords according to the text description of the second content in step 520 is similar to the specific implementation process of extracting the first sub-keywords according to the text description of the first content in step 510. The difference is that step 520 extracts the second sub-keywords according to the text description of the second content, while step 510 extracts the first sub-keywords according to the text description of the first content, and the text descriptions used to extract the keywords are different. To save space, it will not be repeated here.

[0215] In step 530, when the first content is in the form of a video, the first content often includes multiple different modes such as multiple video modes, audio modes, image modes, etc., and the keywords of the first content in each mode may be different. If the keywords in a certain mode are directly extracted alone, it will often cause problems such as inaccurate keywords. Therefore, the embodiment of the present disclosure considers the solution of extracting keywords by multimodal analysis of the first content. Specifically, the specific process of extracting the third sub-keyword based on the multimodal analysis of the first content in the embodiment of the present disclosure includes but is not limited to the following steps:

[0216] Performing video frame extraction on the first content to obtain video frames, and performing feature extraction based on the video frames to obtain visual features;

[0217] Resampling the audio of the first content to obtain preliminary audio, and extracting features based on the preliminary audio to obtain audio features;

[0218] Keywords are extracted based on visual features and audio features to obtain third-level keywords.

[0219] In the specific implementation of this embodiment, the first content is first divided into a number of video clips of equal length, and then a predetermined number of video frames are extracted from each video clip, so as to extract the video frames of the first content and obtain video frames. Then, a preset convolutional neural network is used to extract the two-dimensional visual features and three-dimensional visual features of the video frames, and the two-dimensional visual features and the three-dimensional visual features are feature fused to obtain the final visual features. Further, the first content is audio resampled to obtain preliminary audio, wherein the preliminary audio is monophonic audio. Then, the preliminary audio is short-time Fourier transformed to obtain a spectrogram; the spectrogram is calculated using a preset Mel filter group to obtain a Mel spectrum; the Mel spectrum is logarithmized and framed to obtain the final audio sequence. Then, a preset audio extraction network is used to extract features from the audio sequence to obtain audio features. Finally, a keyword extraction is performed based on a preset visual feature comparison table and visual features to obtain visual keywords, and a keyword extraction is performed based on a preset audio feature comparison table and audio features to obtain audio keywords, and the audio keywords and visual keywords are integrated into third sub-keywords.

[0220] The audio feature may be composed of phonemes, which are the smallest speech units divided according to the natural properties of speech, and are analyzed based on the pronunciation actions in the syllables, where one action constitutes one phoneme. The visual feature may include various feature information such as color information, texture information, text information, motion information, etc. in the first content.

[0221] It should be noted that the preset convolutional neural network is constructed based on the ResNet-152 convolutional neural network pre-trained on the ImageNet dataset and the 3D-ResNet convolutional neural network pre-trained on the Kinetics dataset. The convolutional neural network includes a two-dimensional convolutional part and a three-dimensional convolutional part. The two-dimensional convolutional part is composed of the pre-trained ResNet-152 convolutional neural network, and the three-dimensional convolutional part is composed of the pre-trained 3D-ResNet convolutional neural network. The two-dimensional visual features of the video frame are extracted through the two-dimensional convolutional part of the convolutional neural network, and the three-dimensional visual features of the video frame are extracted through the three-dimensional convolutional part of the convolutional neural network.

[0222] It should be noted that the preset audio extraction network is a VGGish network pre-trained on AudioSet. The preset visual feature comparison table is used to indicate the correspondence between visual features and candidate keywords, and the preset audio feature comparison table is used to indicate the correspondence between audio features and candidate keywords.

[0223] In a specific implementation scenario, the first content is first divided into several video clips with a duration of 10s, and then 8 video frames are extracted from each video clip, and all video frames are integrated to obtain a set of video frames. Then, the two-dimensional convolution part of the convolutional neural network is used to extract the two-dimensional visual features of the video frame, and the three-dimensional convolution part of the convolutional neural network is used to extract the three-dimensional visual features of the video frame. The two-dimensional visual features and the three-dimensional visual features are added to obtain visual features. Further, the first content is audio resampled to obtain a preliminary audio with a frequency of 16kHZ. Then, the preliminary audio is subjected to short-time Fourier transform using a 25ms Hann time window and a 10ms frame shift to obtain a spectrogram; the spectrogram is calculated using a 64-order Mel filter group to obtain a Mel spectrum. Then, the Mel spectrum is logarithmized using a 0.01 offset to obtain a stable Mel spectrum, and the stable Mel spectrum is framed with a duration of 0.96s to obtain the final audio sequence. In the audio sequence, each frame contains 64 Mel frequency bands (excluding frame overlap). Then, the preset audio extraction network is used to extract features from the audio sequence to obtain audio features. Finally, the visual feature comparison table is queried for candidate keywords corresponding to the visual features, and the queried candidate keywords are used as visual keywords; the audio feature comparison table is queried for candidate keywords corresponding to the audio features, and the queried candidate keywords are used as audio keywords; the audio keywords and visual keywords are integrated into the third sub-keyword.

[0224] like Figure 6 As shown, when performing multimodal analysis on the first content and extracting the third sub-keywords, the first content in video form is firstly decomposed into multimodal mode. Since the first content in video form includes audio mode and video mode, the first content is decomposed into two parts according to the different modes, and the audio part of the first content and the video part of the first content are obtained. Next, keyword extraction is performed on the audio part of the first content to obtain four keywords corresponding to the audio part of the first content, and the four keywords are keyword 1, keyword 2, keyword 3 and keyword 4. Keyword extraction is performed on the video part of the first content to obtain four keywords corresponding to the video part of the first content, and the four keywords are keyword 2, keyword 4, keyword 5 and keyword 6. Finally, all keywords are integrated to obtain the third sub-keywords of the first content. Among them, there are 6 third sub-keywords, namely keyword 1, keyword 2, keyword 3, keyword 4, keyword 5 and keyword 6.

[0225] In step 540, when the first sub-keyword, the second sub-keyword, and the third sub-keyword are integrated into the first keyword, first, all the first sub-keywords, the second sub-keywords, and the third sub-keywords are included in the same set. Then, the keywords that appear repeatedly in the set are deduplicated, and multiple identical keywords are deleted so that each keyword appears only once, thereby obtaining the final first keyword.

[0226] It should be noted that each first keyword has a first weight, and the first weight of the first keyword is a fixed weight, which is set according to actual needs.

[0227] Through the above steps 510-540, the embodiment of the present disclosure combines the first content and the text description of the second content placed by the first object before the first content, and uses machine learning and other methods to extract keywords, which can improve the degree of automation of keyword extraction, effectively increase the number of keywords, and improve the accuracy and rationality of the first keyword. In addition, the embodiment of the present disclosure uses a multimodal analysis method to extract keywords for multimodal first content such as video, which can extract keywords in different modes of the first content, achieving multi-dimensionality. Keyword extraction under multimodality can improve the richness and accuracy of the first keyword.

[0228] When the first object is a relatively new author, the second content that only relies on the first content and the first object before the first content often has the problem of a small number of keywords. If the number of keywords is small, sufficient coverage may not be achieved during keyword matching, which will affect the accuracy of keyword matching. The disclosed embodiment considers a solution for keyword expansion based on the first keyword, which can increase the number of keywords, improve the richness of keywords, and thus improve the accuracy of keyword matching.

[0229] Combine the following Figure 7 and Fig. 8A -B describes in detail the process of expanding the second keyword based on the first keyword.

[0230] In some embodiments, the process of expanding the second keyword based on the first keyword includes but is not limited to the following steps 710 to 730 .

[0231] Step 710: input the first keyword into the associated word prediction model to obtain a fourth sub-keyword;

[0232] Step 720: Expand the first keyword using the knowledge graph to obtain a fifth sub-keyword;

[0233] Step 730: Integrate the fourth sub-keyword and the fifth sub-keyword into the second keyword.

[0234] Steps 710 - 730 are described in detail below.

[0235] In step 710, the associated word prediction model can be a model built based on large language models such as GPT-3, T5, BERT, etc., and the associated word prediction model has autonomous learning capabilities. When training the associated word prediction model, first, a number of sample words are input into the associated word prediction model, and the associated word prediction model performs semantic analysis and word prediction on the sample words, learns all the knowledge points contained in the sample words, and generates predicted associated words for each sample word. Then, the model parameters of the associated word prediction model are adjusted according to the semantic proximity between the predicted associated words and the sample words to update the associated word prediction model. After the associated word prediction model is trained, the first associated word is input into the associated word prediction model for semantic analysis and word prediction, and the associated words corresponding to each first associated word are generated, and the associated words are used as the fourth sub-keywords. Among them, each fourth sub-keyword has a confidence probability, and the confidence probability is used to reflect the semantic proximity between the fourth sub-keyword and the first keyword.

[0236] like Fig. 8A As shown, firstly, the first keyword A is input into the associated word prediction model, and the associated word prediction model is used to perform semantic analysis and word prediction on the first keyword A, and three associated words with semantic information close to the first keyword A are obtained. Among them, the three associated words are associated word a1, associated word a2, and associated word a3. The confidence probability of associated word a1 is confidence probability 1, the confidence probability of associated word a2 is confidence probability 2, and the confidence probability of associated word a3 is confidence probability 3. Finally, associated words a1, associated words a2, and associated words a3 are used as the fourth sub-keywords.

[0237] In step 720, when the first keyword is expanded using the knowledge graph, the position of the first keyword in the knowledge graph is first determined; then, based on the position, a first path passing through the first keyword is searched in the knowledge graph, wherein the two endpoints of the first path are a certain word node in the top layer of the knowledge graph and a certain word node in the bottom layer of the knowledge graph. Finally, the words corresponding to each node on the first path are used as the fifth sub-keyword.

[0238] like Figure 8BAs shown, the knowledge graph is divided into four levels. The first level is entertainment, the second level is music series and sports series belonging to entertainment, the third level is Chinese and European and American series belonging to the music series, and tennis, basketball and football series belonging to the sports series. The fourth level is pop and rock styles belonging to Chinese and rock, jazz and pop styles belonging to Europe and America. Assuming that the first keyword is Chinese, when the first keyword is expanded using the knowledge graph, the position of the first keyword in the knowledge graph is first determined. Then, starting from the first keyword "Chinese", search upward or downward to determine the first path passing through the first keyword. Among them, the first path includes "pop-Chinese-music series-entertainment", "rock-Chinese-music series-entertainment". Finally, the words on the first path are used as the fifth sub-keywords, and the fifth sub-keywords obtained are "Chinese-popular", "Chinese-rock" and "music series", "entertainment".

[0239] In step 730, the specific process of integrating the fourth sub-keyword and the fifth sub-keyword into the second keyword is similar to the specific implementation process of integrating the first sub-keyword, the second sub-keyword, and the third sub-keyword into the first keyword in step 540. To save space, it will not be repeated.

[0240] In some embodiments, after the fourth sub-keyword and the fifth sub-keyword are integrated into the second keyword, the content delivery method of the embodiment of the present disclosure further includes determining a first weight of the second keyword, specifically including but not limited to the following steps:

[0241] The confidence probability of the fourth keyword output by the associated word prediction model is used as the first weight of the fourth keyword;

[0242] Determine a first weight of the fifth sub-keyword based on the distance between the fifth sub-keyword and the first sub-keyword in the knowledge graph;

[0243] Based on the first weight of the fourth sub-keyword and the first weight of the fifth sub-keyword, the first weight of the second keyword is determined.

[0244] In the specific implementation of this embodiment, after the first keyword is input into the associated word prediction model, the fourth sub-keyword output by the associated word prediction model will have a confidence probability, which can reflect the semantic proximity between the fourth sub-keyword and the first keyword. Therefore, the confidence probability of the fourth sub-keyword output by the associated word prediction model is directly used as the first weight of the fourth sub-keyword. Due to the distance between the fifth sub-keyword and the first keyword, the semantic proximity between the fifth sub-keyword and the first keyword can be reflected. When the distance between the fifth sub-keyword and the first keyword is closer and the number of levels is less, the semantic proximity between the fifth sub-keyword and the first keyword is higher; when the distance between the fifth sub-keyword and the first keyword is farther and the number of levels is more, the semantic proximity between the fifth sub-keyword and the first keyword is lower. Therefore, according to the distance between the fifth sub-keyword and the first keyword, the first weight of the fifth sub-keyword that is closer to the first keyword is set to a larger value, and the first weight of the fifth sub-keyword that is farther from the first keyword is set to a smaller value. After determining the first weight of the fourth sub-keyword and the first weight of the fifth sub-keyword, the fourth sub-keyword and the fifth sub-keyword are aggregated into the second keyword, and the first weights of the fourth sub-keyword and the fifth sub-keyword are used as the first weight of the second keyword.

[0245] Through the above steps 710-730, the embodiment of the present disclosure adopts a method of combining knowledge graph and keyword prediction model to achieve keyword expansion, which can effectively increase the number of keywords, improve the richness of keywords, and thus improve the accuracy of keyword matching.

[0246] It should be noted that, although the embodiment of the present disclosure shows the implementation process of keyword expansion by combining the knowledge graph and the keyword prediction model, in actual application scenarios, in order to improve the efficiency of keyword expansion, it is also an optional implementation method to use the knowledge graph alone or the keyword prediction model alone for keyword expansion, and its specific implementation process refers to the above steps 710 and 720. To save space, it will not be repeated.

[0247] In some embodiments, when integrating the first keyword and the second keyword as the target keyword, a table template may be set first, the table header of which includes the target keyword and the first weight corresponding to each target keyword. Then, the first keyword, the second keyword and the corresponding first weight are filled into the table template, thereby integrating the first keyword and the second keyword, and obtaining a keyword list consisting of the target keyword and the first weight of each target keyword. Fig. 9As shown in the figure, the keyword list contains 6 target keywords, namely keyword A with a first weight of 0.05, keyword B with a first weight of 0.15, keyword C with a first weight of 0.45, keyword D with a first weight of 0.1, keyword E with a first weight of 0.2, and keyword F with a first weight of 0.05. This method can integrate the first keyword and the second keyword as target keywords in a template filling manner, and each target keyword and its first weight can be more clearly represented through the keyword list, which is conducive to the first object to perform visualization operations on the target keywords.

[0248] It should be noted that since the first keyword is directly extracted based on the first content and the text description of the second content released before the first content is released, and the second keyword is obtained by expanding based on the first keyword, the importance of the first keyword will be higher than that of the second keyword. Therefore, the first weight of the first keyword is often higher than the first weight of the second keyword.

[0249] Since the target keywords generated by the above-mentioned automatic method often do not meet the actual requirements of the first object, if the target keywords are directly used in the subsequent content delivery process, it will often cause problems such as inaccurate content delivery. Based on this, the embodiment of the present disclosure also considers a solution for adjusting the target keywords, so that the first object can adjust the target keywords more conveniently, meet the personalized settings of the target keywords, and improve the flexibility and accuracy of the target keywords.

[0250] In some embodiments, Fig.10 As shown, after integrating the first keyword and the second keyword as the target keyword, the content delivery method of the embodiment of the present disclosure further includes but is not limited to the following steps 1010 to 1020.

[0251] Step 1010: display target keywords;

[0252] Step 1020: Receive the adjustment of the target keyword by the first object.

[0253] Steps 1010 - 1020 are described in detail below.

[0254] In step 1010, after integrating the first keyword and the second keyword into a target keyword, the object terminal displays the target keyword on the front-end interface so that the first object can view the generated target keyword and adjust the target keyword.

[0255] In step 1020, the first object may adjust the target keyword displayed on the front-end interface according to actual needs. The adjustment process includes operations such as modifying the target keyword, adding a target keyword, or deleting the target keyword. After the first object adjusts the target keyword, the object terminal may receive the adjustment of the target keyword by the first object and determine the final target keyword based on the adjustment operation of the first object.

[0256] Through the above steps 1010-1020, the embodiment of the present disclosure can more conveniently display the target keyword to the first object, and the first object can adjust the target keyword according to actual needs, thereby realizing visual editing of the target keyword and satisfying personalized settings of the target keyword, thereby improving the flexibility and accuracy of the target keyword, and further improving the accuracy of matching the target object based on the target keyword and improving the accuracy of content delivery.

[0257] Detailed description of step 330

[0258] In step 330, the object tag of each object in the object set on the content delivery platform is obtained.

[0259] In the specific implementation of this embodiment, since the object label can more clearly reflect the interest preference of each object in the content, when the target object for delivering the first content is screened out from the content delivery object set, the final target object can be determined based on the degree of match between the object label and the target keyword of the first content. Generally speaking, the object label of each object is determined based on a series of object behaviors made by the object to the content on the content delivery platform. The content delivery platform will record the behavior data of each object in the background. Therefore, in the embodiment of the present disclosure, the object label can be obtained based on the content and the behavior of each object to the content. In order to improve the reusability of object labels and improve the efficiency of content delivery, after obtaining the object label, the content delivery platform can also store the object label of each object in a preset object label library for easy access at any time.

[0260] It should be noted that in the embodiments of the present disclosure, when obtaining the object tags of each object in the object set on the content delivery platform, a separate permission or separate consent of the first object will be obtained through a pop-up window or by jumping to a confirmation page. After the separate permission or separate consent of the first object is clearly obtained, the object tags of each object in the object set on the content delivery platform are obtained.

[0261] Combine the following Fig.11 and Fig.12 The process of obtaining the object label of each object is described in detail.

[0262] In some embodiments, the process of obtaining the object tag of each object in the object set on the content delivery platform includes but is not limited to the following steps 1110 to 1140.

[0263] Step 1110: for each object in the object set, obtain a set of content that the object represents interest;

[0264] Step 1120: Obtain the third keyword of the content in the content set;

[0265] Step 1130: for each third keyword, determine the number of contents in the content set that have the third keyword;

[0266] Step 1140: If the number of contents is greater than a fourth threshold, use the third keyword as an object tag of the object.

[0267] Steps 1110 - 1140 are described in detail below.

[0268] In step 1110, since the object's operations such as liking, collecting, and forwarding the published content can be used to determine whether the object is interested in the published content on the content delivery platform, when the object collects a certain content, it indicates that the object is interested in the content. In addition, the application log of the content delivery platform often records the object behavior data of the object's collection, liking, and forwarding of the published content on the content delivery platform in the past. Based on this, for each object in the object set, the object terminal can send a request to obtain the content of interest to the content delivery platform. Then, the content delivery platform will extract the content in which the object has collected, liked, forwarded, etc. from the application log according to the request to obtain the content of interest, and put the extracted series of content into the same set to obtain the content set that the object is interested in. Further, the content delivery platform feeds back the content set that each object is interested in to the object terminal together with the response, so that the object terminal obtains the content set that each object is interested in.

[0269] In step 1120, the specific implementation process of obtaining the third keyword of the content in the content set is similar to the process of extracting the first sub-keyword from the text description of the first content in step 510, except that step 1120 extracts the corresponding third keyword for each content in the content set, while step 510 extracts the corresponding first sub-keyword for the first content. To save space, it will not be repeated here.

[0270] In step 1130, for each object, after obtaining the third keyword of each content in the content set of the object, based on the third keyword, traverse each content in the content set, count the total number of contents with the third keyword, and obtain the number of contents with the third keyword in the content set.

[0271] For example, the content set of object A contains 5 contents, namely, content 1, content 2, content 3, content 4 and content 5, wherein the third keyword of content 1 is keyword M, then the 5 contents of the content set are traversed to check whether the 5 contents contain keyword M. Further, after the traversal operation, it is found that the contents containing keyword M among the 5 contents are content 1, content 3 and content 4, and it is determined that the number of contents with the third keyword in the content set is 3.

[0272] In step 1140, since the number of contents of each third keyword can be used to reflect the importance of the third keyword in the content that the object identifier is interested in, the object label of the object can be determined based on the number of contents of the third keyword. Based on this, the embodiment of the present disclosure adopts a threshold comparison method to screen the third keyword as the object label of the object. Specifically, the number of contents of the third keyword is compared with the fourth threshold. If the number of contents of the third keyword is greater than the fourth threshold, it indicates that the third keyword is a relatively important word in the content that the object identifier is interested in, and therefore, the third keyword is used as the object label of the object. If the number of contents of the third keyword is not greater than the fourth threshold, it indicates that the third keyword is a less important word in the content that the object identifier is interested in, and therefore, the third keyword should not be used as the object label of the object.

[0273] It should be noted that the fourth threshold mentioned above can be set according to actual needs without limitation.

[0274] In order to facilitate the matching of target keywords and object tags in the subsequent keyword matching process, each object tag also has a second weight. Since an object may have one or more object tags, and the frequency of occurrence of each object tag in the content that the object represents interest is different. If the weight of the object tag is set randomly or the weight of each object tag is set to a fixed value, it is often impossible to more truly reflect the importance of each object tag. Therefore, the embodiment of the present disclosure also considers a scheme for determining the second weight of each object tag based on the number of contents of the third keyword corresponding to each object tag, which can more clearly show the importance of each object tag, so as to improve the matching accuracy of the object tag and the target keyword.

[0275] In some embodiments, the process of determining the second weight of each object tag based on the number of contents of the third keyword corresponding to each object tag includes but is not limited to the following steps:

[0276] For each object tag, the number of contents of the third keyword corresponding to the object tag is taken as the number of occurrences of the object tag;

[0277] Count the number of occurrences of all object labels to get the total number of label occurrences;

[0278] For each object tag, the number of occurrences of the object tag and the total number of occurrences of the tag are divided to obtain the second weight of the object tag.

[0279] In the specific implementation of this example, the number of contents of the third keyword corresponding to the object tag is first obtained, and the number of contents is used as the number of occurrences of the object tag. Then, the number of occurrences of all object tags is added to obtain the total number of tag occurrences. Finally, the number of occurrences of each object tag is divided by the total number of tag occurrences to obtain the second weight of the object tag.

[0280] For example, the object tags of object A include tag 1, tag 2, and tag 3, wherein the number of contents of the third keyword 1 corresponding to tag 1 is 4, the number of contents of the third keyword 2 corresponding to tag 2 is 5, and the number of contents of the third keyword 3 corresponding to tag 3 is 11. Based on this, it is determined that the number of occurrences of tag 1 is 4, the number of occurrences of tag 2 is 5, and the number of occurrences of tag 3 is 11. Further, the number of occurrences of these three tags is added together to obtain the total number of times the tags appear, which is 4+5+11=20. Finally, the number of occurrences of each tag is divided by the total number of times the tags appear to obtain the second weight of each tag, wherein the second weight of tag 1 is 0.25, the second weight of tag 2 is 0.2, and the second weight of tag 3 is 0.55.

[0281] like Fig.12 As shown, it is a schematic diagram of the object label list generated according to the object labels of each object in the object set on the content delivery platform. The object label column contains 4 objects, namely, object 1, object 2, object 3, and object 4. Among them, object 1 contains 3 labels, namely, label 1 with a second weight of 0.2, label 2 with a second weight of 0.5, and label 3 with a second weight of 0.3. Object 2 contains 4 labels, namely, label 1 with a second weight of 0.2, label 3 with a second weight of 0.5, label 4 with a second weight of 0.15, and label 5 with a second weight of 0.15. Object 3 contains 2 labels, namely, label 4 with a second weight of 0.7, and label 5 with a second weight of 0.3. Object 4 contains 4 labels, namely, label 11 with a second weight of 0.22, label 13 with a second weight of 0.28, label 10 with a second weight of 0.45, and label 5 with a second weight of 0.05.

[0282] Through the above steps 1110-1140, the embodiment of the present disclosure can more conveniently determine the content of interest represented by each object based on the object behavior data of the objects in the object set, and extract the third keyword of each content of interest in an automated manner. Further, the third keyword that can be used as the object tag of the object is screened out by threshold comparison. In addition, the embodiment of the present disclosure also determines the importance of the object tag based on the number of occurrences of each object tag in the content set, and sets the second weight of the object tag based on the number of occurrences of each object tag in the content set. This method can better improve the accuracy and rationality of the acquired object tags.

[0283] Detailed description of step 340

[0284] In step 340 , a target object is selected from the object set based on the matching degree between the target keyword and the object tag.

[0285] In the specific implementation of this embodiment, since the target keyword has a first weight and the object tag has a second weight, the target keyword and the object tag belong to the same vocabulary system, the first weight and the second weight can be used to calculate the matching degree between the target keyword and the object tag. Further, a number of target objects are selected from the object set by random extraction or by sequential extraction according to the order of matching degree. This method improves the calculation efficiency and accuracy of the matching degree, and can also improve the screening efficiency and accuracy of the target objects.

[0286] like Fig.13 As shown, in some embodiments, the target keyword has a first weight, the object tag has a second weight, and step 340 includes but is not limited to the following steps 1310 to 1320.

[0287] Step 1310: Determine the matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag, and the second weight;

[0288] Step 1320: Select a target object from the object set based on the matching degree of each object in the object set.

[0289] Steps 1310 - 1320 are described in detail below.

[0290] In step 1310, based on the target keyword, the first weight, the object tag, and the second weight, the process of determining the matching degree between the target keyword and the object tag includes but is not limited to the following steps:

[0291] For each object in the object set, determining a target object tag in the object tag of the object that matches one of the target keywords;

[0292] For each target object tag, determine a product of a second weight of the target object tag and a first weight of the matching target keyword;

[0293] The products of each target object label are accumulated to get the matching degree.

[0294] When this embodiment is specifically implemented, first, for each object in the object set, the object tag of the object is compared with the target keyword to determine whether the object tag is consistent with the target keyword. If the object tag is consistent with the target keyword, it is considered that the object tag matches the target keyword, and the object tag that matches the target keyword is used as the target object tag. Then, for each target object tag, the second weight of the target object tag is multiplied by the first weight of the target keyword to obtain the product of the target object tag. Finally, for each object, the products of all target object tags are cumulatively summed, and the result of the cumulative sum is used as the matching degree of the object.

[0295] Specifically, the kth first content item k The first weight of the jth target keyword is w j , for the i-th object u in the object set i , object u i and the jth target keyword t j The second weight of the matched target object label is p ij . Object u i and the kth first content item k The matching degree S(u i ,item k ) can be expressed as shown in formula (1).

[0296]

[0297] In step 1320, based on the matching degree of each object in the object set, several objects can be randomly selected as target objects; or, based on the matching degree, several objects with higher matching degrees or several objects with matching degrees higher than a certain threshold can be selected as target objects, so that the target objects selected from the object set are more in line with expectations.

[0298] Through the above steps 1310-1320, the disclosed embodiment can more conveniently filter out the target object tags that match the target keywords in the object tags of each object, and calculate the matching degree between the object and the first content based on the target object tags and the target keywords, which can improve the accuracy and rationality of the matching degree calculation. Then, the target object is selected from the object set according to the size of the matching degree, which can improve the accuracy of the target object selection, so that the selected target object is more in line with the delivery requirements of the first content, which can improve the accuracy of content delivery.

[0299] Since the object tags of each object in the object set of the content delivery platform can often be obtained through multiple channels, that is, object tags often come from multiple sources. If only object tags from one channel are selected to match the target keywords, there is often a problem of low accuracy in keyword matching. In the embodiment of the present disclosure, a solution for keyword matching based on object tags from multiple sources is provided, which can improve the calculation accuracy of the matching degree.

[0300] like Fig.14 As shown, in some embodiments, the second weight includes a first sub-weight from a first source and a second sub-weight from a second source, and step 1310 includes but is not limited to steps 1410 to 1430 below.

[0301] Step 1410: Determine a first matching degree corresponding to a first source based on a target keyword, a first weight, an object tag, and a first sub-weight;

[0302] Step 1420: Determine a second matching degree corresponding to a second source based on the target keyword, the first weight, the object tag, and the second sub-weight;

[0303] Step 1430: Determine the degree of matching based on the first degree of matching and the second degree of matching.

[0304] Steps 1410-1430 are described in detail below.

[0305] In step 1410, the first sub-weight is the weight of the object tag from the first source. Specifically, first, for each object in the object set, a target object tag matching one of the target keywords among the object tags from the first source is determined. Next, for each target object tag, the first sub-weight of the target object tag is multiplied by the first weight of the matching target keyword to obtain a product result. Finally, the product results of all the target object tags from the first source are accumulated to obtain a first matching degree corresponding to the first source.

[0306] In step 1420, the second sub-weight is the weight of the object tag from the second source. Specifically, first, for each object in the object set, a target object tag matching one of the target keywords among the object tags from the second source is determined. Next, for each target object tag, the second sub-weight of the target object tag is multiplied by the first weight of the matching target keyword to obtain a product result. Finally, the product results of all the target object tags from the second source are accumulated to obtain a second matching degree corresponding to the second source.

[0307] The specific implementation process of the above step 1410 and step 1420 is similar to the specific process of the above step 1310. To save space, they are not repeated here.

[0308] In step 1430, the first matching degree and the second matching degree may be weighted and calculated to obtain the matching degree. Specifically, firstly, a first matching weight corresponding to the first matching degree and a second matching weight corresponding to the second matching degree are set; then, the first matching weight and the first matching degree are multiplied to obtain a first product result, and then the second matching weight and the second matching degree are multiplied to obtain a second product result; finally, the first product result and the second product result are added to obtain the matching degree.

[0309] It should be noted that the sum of the first matching weight and the second matching weight is 1, and the first matching weight and the second matching weight may be equal or unequal without limitation.

[0310] In a specific scenario, the first source is a public account, and the second source is an object video account. First, perform label extraction on the object portrait from the first source to obtain the object's label 1 and the weight of label 1 is 0.7, and the label 2 and the weight of label 2 are 0.3; perform label extraction on the object portrait from the second source to obtain the object's label 2 and the weight of label 2 is 0.4, the label 3 and the weight of label 3 is 0.5, and the label 4 and the weight of label 4 is 0.1. Then, since the downstream application is a video application, the adjustment weight of the first source is set to 0.3 and the adjustment weight of the second source is set to 0.7 according to the correlation between the first source and the second source and the downstream application, and each label is multiplied by the corresponding adjustment weight to obtain the new weight of each label. Among them, for the first source, the new weight of label 1 is 0.21, and the new weight of label 2 is 0.09; for the second source, the new weight of label 2 is 0.28, the new weight of label 3 is 0.35, and the new weight of label 4 is 0.07. Next, sum the weights of all labels to get the total weight, where the total weight is 0.21+0.09+0.28+0.35+0.07=1. Further, divide the weights of all labels by the total weight to get the first sub-weight of each label from the first source and the second sub-weight from the second source. Among them, the first sub-weight of label 1 is 0.21, and the second sub-weight is 0; the first sub-weight of label 2 is 0.09, and the second sub-weight is 0.28; the first sub-weight of label 3 is 0, and the second sub-weight is 0.35; the first sub-weight of label 4 is 0, and the second sub-weight is 0.07.

[0311] Through the above steps 1410-1430, the embodiment of the present disclosure can assign different weights to object tags from different sources, and then match object tags from multiple channels with target keywords, which can effectively improve the calculation accuracy of the matching degree.

[0312] Since randomly selecting several candidate objects as target objects from the object set according to the matching degree often results in the selected target objects being inaccurate, the target objects may not be the objects that are more interested in the first content, resulting in a high degree of deviation in the first content reaching the target objects of interest, thus affecting the delivery accuracy of the first content. The disclosed embodiment considers a solution for selecting target objects from the object set by using threshold comparison and sorting selection, which can effectively improve the accuracy of selecting target objects, enable the first content to reach more target objects of interest, and thus improve the delivery accuracy of the first content.

[0313] Combine the following Fig.15 and Fig.16A -C describes the specific process of selecting target objects from an object set by using threshold comparison and sorting selection.

[0314] In one embodiment, the target object can be screened out from the object set based only on the threshold comparison method. Specifically, first, the candidate objects whose matching degree is greater than the third threshold are determined. Then, the candidate objects whose matching degree is greater than the third threshold are used as the target objects. Fig.16A As shown, the object set includes 6 objects, namely object 1, object 2, object 3, object 4, object 5 and object 6. Among them, the matching degree of object 1 is 0.88, the matching degree of object 2 is 0.96, the matching degree of object 3 is 0.73, the matching degree of object 4 is 0.42, the matching degree of object 5 is 0.2, and the matching degree of object 6 is 0.68. The third threshold is set to 0.5, and the candidate objects with matching degrees higher than the third threshold are selected as target objects. Then the target objects include object 1, object 2, object 3 and object 6. This embodiment uses the threshold comparison method to filter out the target objects from the object set, which can clearly and easily implement object screening and improve the efficiency and accuracy of object screening.

[0315] In another embodiment, the target object can be selected from the object set based on the sorting selection method. Specifically, the objects in the object set are first sorted in descending order of matching degree; then, the object in the first position in the sorting is used as the target object. Fig. 16BAs shown, the object set contains 6 objects, namely object 1, object 2, object 3, object 4, object 5 and object 6. Among them, the matching degree of object 1 is 0.88, the matching degree of object 2 is 0.96, the matching degree of object 3 is 0.73, the matching degree of object 4 is 0.42, the matching degree of object 5 is 0.2, and the matching degree of object 6 is 0.68. According to the order of matching degree from high to low, the six objects are sorted, and the order is object 2, object 1, object 3, object 6, object 4, object 5. The objects ranked in the top three are selected as target objects, and the target objects include object 2, object 1 and object 3. This embodiment uses the sorting method to select to filter out the target object from the object set, which can improve the efficiency and rationality of object screening.

[0316] like Fig.15 As shown, in another embodiment, based on the matching degree of each object in the object set, the process of selecting the target object from the object set includes but is not limited to the following steps 1510 to 1530.

[0317] Step 1510: Determine candidate objects in the object set whose matching degree is greater than a third threshold;

[0318] Step 1520: sort the candidate objects in descending order of matching degree;

[0319] Step 1530: Take the candidate object in the first position in the sorting as the target object.

[0320] Steps 1510-1530 are described in detail below.

[0321] In step 1510, since the higher the matching degree of the object, the more likely the object is to be interested in the first content, the object is more suitable as the delivery object of the first content. First, a third threshold is set, and then the matching degree of each object in the object set is compared with the third threshold. Finally, the object with a matching degree greater than the third threshold is selected as a candidate object.

[0322] In step 1520, according to the height order of the matching degrees of the candidate objects, the candidate objects are sorted in order from high to low matching degrees.

[0323] In step 1530, since the higher the matching degree, the higher the position of the candidate object, the first candidate object in the ranking is used as the target object, wherein the first position can be set according to actual needs without limitation.

[0324] like Fig. 16CAs shown, the object set contains 6 objects, namely object 1, object 2, object 3, object 4, object 5 and object 6. Among them, the matching degree of object 1 is 0.88, the matching degree of object 2 is 0.96, the matching degree of object 3 is 0.73, the matching degree of object 4 is 0.42, the matching degree of object 5 is 0.2, and the matching degree of object 6 is 0.68. First, the third threshold is set to 0.4, and the candidate objects with matching degrees higher than the third threshold are selected as candidate objects, and the candidate objects include object 1, object 2, object 3, object 4 and object 6. Then, the five candidate objects are sorted in order from high to low matching degrees, and the order is object 2, object 1, object 3, object 6, object 4. The top three objects are selected as target objects, and the target objects include object 2, object 1 and object 3.

[0325] Through the above steps 1510-1530, the embodiment of the present disclosure combines the two methods of threshold comparison and sorting selection to filter out target objects from the object set. Compared with the method of using only threshold comparison for object screening or the method of using only sorting selection for object screening, this method can significantly improve the screening accuracy of target objects, so that the first content can reach more target objects of interest, thereby improving the delivery accuracy of the first content.

[0326] Since only using keyword matching to screen out target objects often makes the first object unable to adjust the target objects according to actual needs, it is not conducive to selecting personalized settings of target objects from the object set, and will affect the accuracy of selecting target objects. The disclosed embodiment considers a solution for screening target objects based on a combination of keyword matching and crowd screening, which can realize the personalization of screening target objects, make the selected target objects more in line with the actual needs of the first object, and also improve the accuracy and rationality of selecting target objects.

[0327] In some embodiments, Fig.17 As shown, the process of screening out target objects from an object set based on keyword matching and crowd screening includes but is not limited to the following steps 1710 to 1730 .

[0328] Step 1710: Based on the matching degree between the target keyword and the object label, perform a first screening in the object set to obtain a first screened object;

[0329] Step 1720: Obtain a first screening condition of the first object;

[0330] Step 1730: Based on the matching of the first screening condition and the attribute information of the first screened object, a second screening is performed in the first screened object to obtain the target object.

[0331] Steps 1710-1730 are described in detail below.

[0332] In step 1710, based on the matching degree between the target keyword and the object tag, the first screening is performed in the object set, and the specific process of obtaining the first screened object is similar to the specific implementation process of the above step 1320. The difference is that step 1710 performs the first screening on the object set based on the matching degree to obtain the first screened object; while step 1320 screens the target object from the object set based on the matching degree. To save space, it will not be repeated.

[0333] In step 1720, the editing interface corresponding to the first filtering condition is first displayed to the first object, and the first object enters the specific content of the first filtering condition on the editing interface. Then, the object terminal can receive the first filtering condition entered by the first object. The specific content of the first filtering condition includes filtering conditions for tags, filtering conditions for the activity of the crowd, filtering conditions for the region of the crowd, filtering conditions for the education level of the crowd, filtering conditions for the number of years of account registration, etc., without limitation.

[0334] In step 1730, the attribute information of the first filtered object refers to information such as the activity, location, object label, education level, and account registration years of the first filtered object. First, the attribute information of the object is obtained from the log data of each first filtered object in the content delivery platform. Then, when the first filtered object is subjected to a second filter based on the match between the first filtering condition and the attribute information of the first filtered object, each attribute information of the first filtered object is compared with the specific content in the first filtering condition for the first filtered object; if all the attribute information of the first filtered object meets the specific content of the first filtering condition, the first filtered object is used as the target object.

[0335] For example, the first object sets the first screening condition as "activity is not less than medium activity, and the account registration period is greater than 1 year". There are 5 objects after the first screening, and the attribute information of the objects after the first screening includes activity and account registration period. These 5 objects after the first screening are object 1 with high activity and account registration period of 8 months, object 2 with high activity and account registration period of 10 years, object 3 with medium activity and account registration period of 5 years, object 4 with low activity and account registration period of 1 month, and object 5 with low activity and account registration period of 2 years. Comparing the attribute information of these five objects after the first screening with the first screening condition, it is found that: the account registration period of object 1 does not meet the requirements, the activity and account registration period of object 4 do not meet the requirements, the activity of object 5 does not meet the requirements, and the activity and account registration period of objects 2 and 3 meet the requirements. Therefore, objects 2 and 3 in the objects after the first screening are taken as target objects.

[0336] Through the above steps 1710-1730, the embodiment of the present disclosure combines the two methods of keyword matching and crowd screening to select the target object from the object set. Compared with the method of matching with a single indicator or the method of manual selection in the related art, this method can better match the object that meets the delivery requirements. In addition, based on the combination of keyword matching and crowd screening to screen the target object, it is also possible to personalize the screening of the target object, so that the selected target object is more in line with the actual needs of the first object, and it is also possible to improve the accuracy and rationality of the selection of the target object.

[0337] Since the first filtering condition is often set by the first object based on habitual preferences, etc., and is often set when the first object does not understand the overall attribute distribution of the objects after the first filtering, this situation may lead to problems such as a small number of target objects finally filtered out, affecting the scope and breadth of content delivery, and is not conducive to increasing the exposure of the first content. Based on this, the embodiment of the present disclosure considers a solution of displaying the screening results based on the first filtering condition to the first object, so that the first object can more conveniently adjust the filtering condition and further filter out the target object, and can flexibly adjust the filtering condition to make the filtering condition set by the first object more reasonable, thereby making the number of target objects more reasonable, and then making the first content better reach the target object of interest and improve the accuracy of content delivery.

[0338] Combine the following Fig.18 and Fig.19A -B describes in detail the process of performing a second screening on the objects after the first screening to obtain the target objects.

[0339] Step 1810: Based on the matching of the first screening condition and the attribute information of the first screened object, a second screening is performed in the first screened object to obtain a second screened object;

[0340] Step 1820, obtaining attribute information of the second filtered object;

[0341] Step 1830, displaying a distribution diagram of the objects after the second screening based on the attribute information;

[0342] Step 1840: display an attribute information checkbox to receive a second filter condition input by the first object based on the distribution graph;

[0343] Step 1850: Based on the matching of the second screening condition and the attribute information of the second screened object, a third screening is performed on the second screened object to obtain the target object.

[0344] Steps 1810-1850 are described in detail below.

[0345] In step 1810, based on the matching of the first screening condition and the attribute information of the first screened object, the second screening is performed in the first screened object, and the specific implementation process of obtaining the second screened object is similar to the above step 1730. The difference is that the second screening in step 1810 obtains the second screened object, while the second screening in step 1730 obtains the target object. To save space, it will not be repeated.

[0346] In step 1820, since the attribute information of each object is clearly recorded in the log data of the content delivery platform, the object terminal can obtain the attribute information of the object from the log data of each second filtered object in the content delivery platform. Specifically, the object terminal first sends a request to obtain the attribute information to the content delivery platform; then, the content delivery platform retrieves the log data of each second filtered object from the backend server according to the request, and extracts the attribute information of each second filtered object from the log data. Finally, the content delivery platform will feed back the attribute information of each second filtered object together with the response to the object terminal, so that the object terminal can obtain the attribute information of each second filtered object.

[0347] In step 1830, the total number of objects of the second filtered object is first counted. Then, the number of objects corresponding to each attribute item under each attribute information is counted, and the quotient of the number of objects and the total number of objects is calculated, and the proportion of the number of people of each attribute item under each attribute information is determined according to the quotient. Finally, a distribution map of the second filtered object is generated according to the proportion of the number of people of each attribute item under each attribute information, and the distribution map is displayed to the first object on the front-end interface of the object terminal. Among them, the distribution map includes the distribution of the number of objects of each attribute information, and the distribution map includes a histogram, a pie chart, etc., without limitation.

[0348] like Fig.19A As shown, in the attribute distribution diagram of the second screened objects, the total number of the second screened objects is 160, and the attribute distribution diagram includes three parts. The first part is a pie chart reflecting the proportion of activity, among which there are 100 second screened objects with high activity and 60 second screened objects with medium activity. The second part is a pie chart reflecting the account registration years, among which there are 60 second screened objects with an account registration year of no more than 1 year, 80 second screened objects with an account registration year between 1 and 5 years, and 20 second screened objects with an account registration year of more than 5 years. The third part is a histogram reflecting the distribution of object labels, among which there are 32 second screened objects with label 1, 48 second screened objects with label 2, 16 second screened objects with label 3, and 64 second screened objects with label 4. When the first object determines that the attribute distribution of the second screened object meets expectations, click the "OK" button to complete the screening of the target object. When the first object needs to adjust the filtering condition for filtering the target object, the first object can click the "check" button according to the specific situation of the attribute distribution graph to input the second filtering condition based on the distribution graph.

[0349] In step 1840, an attribute information checkbox is displayed on the front-end interface of the object terminal, and the attribute information checkbox is provided to the first object for selection to input the second filtering condition. When the first object selects the attribute information checkbox, the object terminal displays the setting interface of the second filtering condition in response to the selection operation of the first object, and the first object inputs the specific content of the second filtering condition on the setting interface of the second filtering condition. The second filtering condition is used to match the attribute information of the second filtered object, and the third filtering is performed in the second filtered object to obtain the target object.

[0350] like Fig.19BAs shown, after the first object clicks the "check" button, the object terminal displays the setting interface of the second filtering condition, on which the input boxes for each filtering item in the second filtering condition and the prompt field "Please enter the second filtering condition" are displayed, wherein the filtering items include object label, crowd activity, account registration years, and crowd region. The first object enters the fields "Label 1, Label 4" in the input box corresponding to the object label, enters the field "≥ Medium Activity" in the input box corresponding to the crowd activity, enters the field ">1 year" in the input box corresponding to the account registration years, and enters the field "City A" in the input box corresponding to the crowd region. After completing the input of each filtering item, the first object clicks the "OK" button to confirm the completion of the setting of the second filtering condition.

[0351] In step 1850, based on the matching of the second screening condition and the attribute information of the second screened object, the third screening is performed on the second screened object to obtain the target object. The specific implementation process is similar to the above step 1730. The difference is that step 1850 performs the third screening on the second screened object based on the second screening condition to obtain the target object, while step 1730 performs the second screening on the first screened object based on the first screening condition to obtain the target object. To save space, it will not be repeated.

[0352] Through the above steps 1810-1850, the embodiment of the present disclosure displays the distribution of the attribute information of the second filtered objects to the first object, so that the first object can more conveniently adjust the filtering conditions, obtain the second filtering conditions, and further filter out the target objects according to the second filtering conditions. This method can flexibly adjust the filtering conditions, make the filtering conditions set by the first object more reasonable, thereby making the number of target objects more reasonable, and can better control the scope and breadth of content delivery, effectively increase the exposure of the first content, enable the first content to better reach the target objects of interest, and improve the accuracy of content delivery.

[0353] Since the matching degrees of different second-screened objects are different, the object labels they contain will be quite different. When the number of objects after the second screening is large, if the object labels of the second-screened objects are displayed to the first object one by one for verification and screening, it often takes a lot of time, which will result in low efficiency of content delivery. The disclosed embodiment takes into account a scheme for sampling display based on matching degree, which can reasonably display the object labels of the second-screened objects to the first object so that the first object can select the appropriate number of retained people, which can improve the convenience of object display and the rationality of selecting the number of retained people, and can also effectively improve the efficiency of content delivery.

[0354] Combine the following Fig. 20 and Fig.21A-C describes in detail the process of sampling display based on matching degree.

[0355] Step 210: Based on the matching of the screening condition and the attribute information of the first screened object, a second screening is performed in the first screened object to obtain a second screened object;

[0356] Step 220: sort the second screened objects according to the matching degree from high to low;

[0357] Step 2030: taking the second screened object in the top second position in the sorting as the first display object, and displaying the first display object and the object label of the first display object that matches the target keyword;

[0358] Step 2040: Select some of the second screened objects from the second screened objects that are not ranked in the second position in the sorting as second display objects, and display the second display objects and object labels of the second display objects that match the target keyword;

[0359] Step 2050: receiving a reserved number input by the first object;

[0360] Step 2060: retain the second screened objects according to the retention number to obtain the target objects.

[0361] Steps 2010-2060 are described in detail below.

[0362] In step 210, based on the matching of the screening condition with the attribute information of the first screened object, the second screening is performed in the first screened object, and the specific implementation process of obtaining the second screened object is similar to the above step 1730. The difference is that the second screening in step 2010 obtains the second screened object, while the second screening in step 1730 obtains the target object. To save space, it will not be repeated.

[0363] In step 220, according to the matching degree of each second screened object, the second screened objects are arranged in descending order according to the order of matching degree to obtain a ranking of the second screened objects. In the ranking, the second screened objects with higher matching degree are ranked at the front, and the second screened objects with lower matching degree are ranked at the back.

[0364] In step 2030, the specific numerical value of the top second rank can be set according to the actual situation without limitation. Specifically, the specific numerical value of the second rank is first preset. Then, the second screened object ranked in the top second rank is determined in the sorting. Further, the second screened object ranked in the top second rank is used as the first display object. Finally, the object tag that matches the target keyword in the object tag of the first display object is determined, and the first display object and its object tag that matches the target keyword are displayed to the first object on the front-end interface.

[0365] like Fig.21A As shown, the second screened objects ranked in the top three in the sorting are taken as the first display objects, wherein the first display objects include the first display object 1, the first display object 2 and the first display object 3, the matching degree of the first display object 1 is 0.96, the matching degree of the first display object 2 is 0.87, and the matching degree of the first display object 3 is 0.77. Next, it is determined that the object tags matching the target keyword in the object tags of the first display object 1 include tags 1, tags 2 and tags 3, it is determined that the object tags matching the target keyword in the object tags of the first display object 2 include tags 1 and tags 3, and it is determined that the object tags matching the target keyword in the object tags of the first display object 3 include tags 1 and tags 3. Each first display object and its object tags matching the target keyword are displayed to the first object.

[0366] In step 2040, the number of second screened objects selected from the second position that has not been ranked in the sorting is first determined. Then, according to the predetermined number, the number of second screened objects selected from the second position that has not been ranked in the sorting is randomly selected as the second display objects. Finally, the object tags that match the target keyword in the object tags of the second display objects are determined, and the second display objects and their object tags that match the target keyword are displayed to the first object on the front-end interface.

[0367] like Fig.21AAs shown, first determine that the number of second screened objects selected from the second position that has not been ranked in the sorting is 6. Then, randomly select six second screened objects from the second position that has not been ranked in the sorting as second display objects, wherein the second display objects include second display object 1 with a matching degree of 0.5, second display object 2 with a matching degree of 0.42, second display object 6 with a matching degree of 0.45, second display object 4 with a matching degree of 0.62, second display object 5 with a matching degree of 0.33, and second display object 6 with a matching degree of 0.67. Then, determine that the object label of the second display object 1 that matches the target keyword is label 1, determine that the object label of the second display object 2 that matches the target keyword is label 2, determine that the object label of the second display object 3 that matches the target keyword is label 1, determine that the object label of the second display object 4 that matches the target keyword is label 3, determine that the object label of the second display object 5 that matches the target keyword is label 1, and determine that the object label of the second display object 6 that matches the target keyword is label 3. Finally, each second display object and its object label that matches the target keyword are displayed to the first object.

[0368] In step 2050, the reserved number is selected by the first object according to the first display object, the object tag of the first display object matching the target keyword, the second display object, and the object tag of the second display object matching the target keyword. Specifically, the first object inputs the reserved number on the front-end interface according to each first display object, second display object and each object tag displayed. Then, the object terminal receives the reserved number input by the first object. The reserved number is the number of objects of the target object to which the first content will be delivered.

[0369] like Fig. 21B As shown, the target terminal displays an input box for the reserved number and an "OK" button on the interface. The first target enters the reserved number "100" on the interface and clicks the "OK" button. At this time, the target terminal receives the reserved number entered by the first target and determines the target number of the target object to be 100.

[0370] In step 2060, the second screened objects are retained according to the retained number, and specific situations of obtaining the target objects include but are not limited to the following two situations:

[0371] Case 1: According to the reserved number, the second screened object ranked first by the reserved number in the sorting is selected as the target object;

[0372] In case 2, according to the reserved number, the reserved number of second screened objects are randomly selected as the target objects.

[0373] like Fig. 21CAs shown, the number of reservations received by the object terminal is 100. In the ranking determined by the degree of matching, the second screened objects ranked in the top 100 are selected as target objects. The final target object object list is displayed on the front-end interface, and the object list includes the target object and the degree of matching of each target object. The first object clicks the "OK" button to determine the target object to which the first content is to be delivered.

[0374] Through the above steps 2010-2060, the disclosed embodiment displays the second screened objects with a higher matching degree in the top second rank as the first display objects one by one to the first object, which can clearly reflect the specific category of the object label contained in the second screened objects with a higher matching degree. In addition, for the second screened objects whose matching degree is not ranked in the top second rank, a sampling method is adopted to randomly select a part of the second screened objects as the second display objects for display, which can better improve the convenience and rationality of object display, thereby improving the efficiency of content delivery. It can also enable the first object to determine a reasonable number of retentions in combination with the first display object and the second display object, thereby improving the rationality of the number of people delivered.

[0375] Another implementation method of the content delivery method of the disclosed embodiment

[0376] After the first content is delivered to the target object, since the object tags of different target objects are different, the interest expressed by each target object in the first content will be different. If the first weight of the target keyword of the first content is set to a fixed value by default, it is often not conducive to the accurate delivery of the first content. Based on this, the embodiment of the present disclosure considers a solution for dynamically adjusting the first weight of the target keyword based on the different levels of interest of each target object in the first content, which can achieve iterative optimization of the first weight of the target keyword, thereby improving the accuracy of content delivery.

[0377] Combine the following Fig. 22 Another specific implementation process of the content delivery method of the embodiment of the present disclosure is described in detail.

[0378] Step 310: Determine that a first condition is met based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content;

[0379] Step 320: In response to the first condition being met, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword as a target keyword;

[0380] Step 330: Obtain object tags for each object in the object set on the content delivery platform;

[0381] Step 340: Select a target object from the object set based on the matching degree between the target keyword and the object tag;

[0382] Step 350: delivering the first content based on the selected target object;

[0383] Step 2210: Obtain the interest expression of the target object after the first content is delivered;

[0384] Step 2220: Update the first weight of the target keyword of the first content according to the interest expression of the target object.

[0385] Steps 310-2220 are described in detail below.

[0386] Step 310 to step 350 have been described in detail in the previous text. Therefore, for the specific process of step 310 to step 350 in the specific implementation of this embodiment, please refer to the previous description and will not be repeated here.

[0387] It should be noted that, in the embodiment of the present disclosure, when the number of published content of the same type as the first content to be delivered and the number of published content of the first object for delivering the first content are obtained, a separate permission or separate consent of the first object will be obtained through a pop-up window or by jumping to a confirmation page. After the separate permission or separate consent of the first object is clearly obtained, the number of published content of the same type as the first content to be delivered and the number of published content of the first object for delivering the first content are obtained.

[0388] In addition, when obtaining the object tags of each object in the object set on the content delivery platform, the separate permission or separate consent of the first object will be obtained through a pop-up window or jumping to a confirmation page. After clearly obtaining the separate permission or separate consent of the first object, the object tags of each object in the object set on the content delivery platform will be obtained.

[0389] In step 2210, after delivering the first content to the target object of the content delivery platform, the background server of the content delivery platform will record the behavior data of each target object on the first content, wherein the behavior data includes data generated by behaviors such as clicks, favorites, likes, comments, or forwarding. Then, the background server of the content delivery platform will generate an interest expression of the target object based on the behavior data of each target object, and send the interest expression of the target object to the object terminal. Among them, the interest expression refers to whether the target object is interested in the first content. Specifically, when generating the interest expression of the target object based on the behavior data of each target object, a judgment condition can be set first, and then, based on the match between the judgment condition and the behavior data, the interest expression of the target object is expressed as interest or disinterest.

[0390] For example, the judgment condition is set to "If the target object has at least one of the following behaviors of collecting, liking, and forwarding the first content, it is determined that the target object is interested in the first content." Then, based on the behavior data of each target object, if the behavior data contains at least one of the following behaviors of collecting, liking, and forwarding, it is considered that the target object's interest is expressed as interest in the first content. If the behavior data does not contain the following behaviors of collecting, liking, and forwarding, it is considered that the target object's interest is expressed as not being interested in the first content.

[0391] In step 2220, the first weight is the ratio of the first number to the second number, the first number being the number of times the target object expresses interest in the target keyword after the first content with the target keyword is delivered to the target object, and the second number being the number of times the first content with the target keyword is delivered to the target object. Specifically, when updating the first weight of the target keyword of the first content according to the interest expression of the target object, the target keyword in the first content that matches the object tag of the target object whose interest is expressed as being interested in the first content is obtained, and the first weight of the obtained target keyword is adjusted to a higher value. Since the sum of the first weights of all target keywords in the first content is 1, the first weight of the target keyword in the first content that does not match the object tag of the target object whose interest is expressed as being interested in the first content is adjusted to a lower value.

[0392] Through the above steps 310-2220, the embodiment of the present disclosure can more conveniently determine whether the weights of each target keyword of the first content are reasonable based on the target object's interest expression in the first content, and adjust and update the first weight of each target keyword according to the different object labels carried by the target objects with different interest expressions, so as to more conveniently realize the adaptive iterative optimization of the first weight of the target keyword, thereby improving the delivery accuracy of the subsequent content delivery process.

[0393] Another implementation method of the content delivery method of the disclosed embodiment

[0394] After the first content is delivered to the target object, since different target objects have different object tags, the interest expressed by each target object in the first content will be different. If the second weight of the target object's object tag is set to a fixed value by default, it is often not conducive to accurately screening out the target objects interested in the first content, and will cause the first content to fail to reach the target objects of interest well. Based on this, the embodiment of the present disclosure considers a solution for dynamically adjusting the second weight of the object tag based on the different levels of interest of each target object in the first content, which can achieve iterative optimization of the second weight of the object tag, thereby improving the accuracy of the screened target objects, so that the first content can better reach the target objects of interest.

[0395] Combine the following Fig.23 Another specific implementation process of the content delivery method of the embodiment of the present disclosure is described in detail.

[0396] Step 310: Determine that a first condition is met based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content;

[0397] Step 320: In response to the first condition being met, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword as a target keyword;

[0398] Step 330: Obtain object tags for each object in the object set on the content delivery platform;

[0399] Step 340: Select a target object from the object set based on the matching degree between the target keyword and the object tag;

[0400] Step 350: delivering the first content based on the selected target object;

[0401] Step 2310: Obtain the interest expression of the target object after the first content is delivered;

[0402] Step 2320: Update the second weight of the object tag of the target object according to the interest expression of the target object.

[0403] Steps 310-2320 are described in detail below.

[0404] Step 310 to step 350 have been described in detail in the previous text. Therefore, for the specific process of step 310 to step 350 in the specific implementation of this embodiment, please refer to the previous description and will not be repeated here.

[0405] The specific implementation process of step 2310-step 2320 is similar to the implementation process of step 2210-step 2220. The difference is that step 2310-step 2320 is to update the second weight of the object tag of the target object according to the interest expression of the target object, while step 2210-step 2220 is to update the first weight of the target keyword of the first content according to the interest expression of the target object, and the update objects of the two are different. In order to save space, it will not be repeated.

[0406] Among them, the second weight is the ratio of the third number to the fourth number, the third number is the number of times the target object expresses interest after the first content is delivered to the target object with the object tag, and the fourth number is the number of times the first content is delivered to the target object with the object tag.

[0407] Through the above steps 310-2320, the embodiment of the present disclosure can more conveniently determine whether the second weights of each object label of the target object are reasonable according to the target object's interest expression in the first content, and adjust and update the second weights of each object label of the target object according to the differences in the object labels carried by the target objects with different interest expressions, so as to more conveniently realize the adaptive iterative optimization of the second weights of the object labels, thereby improving the accuracy of the screened target objects and enabling the first content to better reach the target objects of interest.

[0408] Another implementation method of the content delivery method of the disclosed embodiment

[0409] After the first content is delivered to the target object, since the attribute information of different target objects is different, the interest expressed by each target object in the first content will be different. In addition, the attribute information can reflect to a certain extent whether the screening conditions set in the content delivery process are reasonable. Based on this, the embodiment of the present disclosure considers a scheme of counting the interest ratio of the target objects under each attribute information, and displaying the counted ratio to the first object as a reference for subsequent delivery tasks, which can improve the content delivery accuracy of each delivery task after the first content is delivered.

[0410] Combine the following Fig.24 , Fig.25 Another specific implementation process of the content delivery method of the embodiment of the present disclosure is described in detail.

[0411] Step 310: Determine that a first condition is met based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content;

[0412] Step 320: In response to the first condition being met, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword as a target keyword;

[0413] Step 330: Obtain object tags for each object in the object set on the content delivery platform;

[0414] Step 340: Select a target object from the object set based on the matching degree between the target keyword and the object tag;

[0415] Step 350: delivering the first content based on the selected target object;

[0416] Step 2410: Obtain the interest expression of the target object after the first content is delivered;

[0417] Step 2420: Obtain attribute information of the target object;

[0418] Step 2430: for each piece of attribute information, determine the ratio of interest in the target object corresponding to the attribute information;

[0419] Step 2440: Display the ratios corresponding to each attribute information.

[0420] Steps 310-2440 are described in detail below.

[0421] Step 310 to step 350 have been described in detail in the previous text. Therefore, for the specific process of step 310 to step 350 in the specific implementation of this embodiment, please refer to the previous description and will not be repeated here.

[0422] In step 2410, the specific implementation process of obtaining the interest expression of the target object after the first content is delivered is similar to the above step 2210. To save space, it will not be repeated.

[0423] In step 2420, the specific implementation process of obtaining the attribute information of the target object is similar to the specific implementation process of obtaining the attribute information of the first filtered object in step 1730. The difference is that step 2420 obtains the attribute information of the target object, while step 1730 obtains the attribute information of the first filtered object. To save space, it will not be repeated.

[0424] In step 2430, first, the total number of objects of all target objects may be counted. Next, for each attribute information, the total number of people who are interested and the total number of people who are not interested in the target object corresponding to the attribute information are counted. Finally, the total number of people who are interested in the target object corresponding to the attribute information is divided by the total number of objects to obtain the ratio of people who are interested in the target object corresponding to the attribute information.

[0425] like Fig.25As shown, the attribute information of the target object includes activity, registration years and object label, and the attribute distribution diagram of the target object includes three parts. The first part is the ratio distribution diagram of activity, in which the ratio of target objects with low activity and no interest in the first content is 12.5%, the ratio of target objects with low activity but interest in the first content is 25%, and the ratio of target objects with high activity and interest in the first content is 62.5%. The second part is the ratio distribution diagram of registration years, in which the ratio of target objects with a registration year of not less than 5 years but no interest in the first content is 12.5%, the ratio of target objects with a registration year between 1 and 5 years but interest in the first content is 37.5%, the ratio of target objects with a registration year between 1 and 5 years but no interest in the first content is 12.5%, and the ratio of target objects with a registration year of not more than 1 year but interest in the first content is 37.5%. The third part is the ratio distribution diagram of object labels, among which, among the target objects with label 1, 10 are interested in the first content, and 10 are not interested in the first content; among the target objects with label 2, 20 are interested in the first content, and 10 are not interested in the first content; among the target objects with label 3, 30 are interested in the first content, and no one is not interested in the first content; among the target objects with label 4, 5 are interested in the first content, and 15 are not interested in the first content.

[0426] In step 2440, after the ratio of interest is indicated in the target object corresponding to the attribute information and the attribute distribution graph is generated, the ratio distribution corresponding to each attribute information is displayed to the first object on the front-end interface. The ratio of each attribute information is used for the delivery screening of the third content of the first object after the first content is delivered, so that the target object corresponding to the third content is screened more accurately, thereby improving the accuracy of content delivery in the subsequent delivery stage.

[0427] Through the above steps 310-2440, the embodiment of the present disclosure can analyze the ratio of interest in the target object for each attribute information and display the ratio to the first object, which can help the first object to screen out more accurate target objects in the screening of the third content after the first content is delivered, thereby improving the efficiency of the third content in reaching the target object of interest, and can also improve the content delivery accuracy of each delivery task after the first content is delivered.

[0428] The specific interaction process between the object and the object terminal in the content delivery method of the embodiment of the present disclosure Refer to the following Fig.26 , the implementation details of the content delivery method of the embodiment of the present disclosure are explained in detail and by way of example.

[0429] After receiving the first content to be delivered uploaded by the first object A, the content delivery recommendation system 141 in the object terminal 140 first extracts keywords based on the first content to obtain the target keywords. For the specific implementation process, refer to the above step 320. Next, obtain the object tags of each object in the object set on the content delivery platform, match the target keywords with the object tags, and screen out the preliminary selected population to be screened from the object set according to the matching situation. For the specific implementation process, refer to the above steps 330 and 1710. Next, screen the preliminary selected population to be screened to obtain the preliminary delivery population and package mining results. For the specific process, refer to the above step 1810. Furthermore, the preliminary delivery population and package mining results are analyzed in the analysis and screening link. The analysis process includes population magnitude analysis, basic attribute analysis, scene activity analysis, object sampling analysis, etc. The analysis result is presented to the first object A, so that the first object A can input new screening conditions or remove the hit objects; then, after the first object A sets new screening conditions, the crowd screening stage is performed again, and the crowd screening is performed again according to the new screening conditions to obtain the final delivery crowd and mining results. The specific implementation process refers to the above steps 1820 to 1850. Finally, the object terminal 140 delivers the first content to the delivery crowd of the content delivery platform in an online delivery manner.

[0430] Furthermore, after the online delivery is completed, the delivery level, the comparison of delivery indicators and the delivery effect are displayed to the first object. Furthermore, the delivery result is connected to the content delivery recommendation system 141, and the first weight of the target keyword and the second weight of the object label are updated according to the interest level of each target object in the delivery result in the first content, so as to achieve iterative optimization of the first weight and the second weight. For the specific implementation process, refer to the above steps 2220 and 2320. At the same time, according to the interest level of each target object in the delivery result in the first content, the interest ratio of the target object under each attribute information is generated, and the ratio is displayed to the first object, and the ratio is used as the first object in the subsequent content delivery process. The crowd screening condition setting basis, the specific implementation process refers to the above steps 2430 to 440.

[0431] Refer to the following Fig. 27 , describing in detail a specific usage process of the content delivery method of an embodiment of the present disclosure, which process includes but is not limited to the following steps 2701 to 2722.

[0432] Step 2701: The first object A sends a content delivery request to the object terminal 140;

[0433] Step 2702: The target terminal 140 obtains the number of published contents of the same type as the first content in the content delivery request and the number of published contents of the first target that delivers the first content;

[0434] Step 2703: The target terminal 140 determines that the first condition is satisfied according to the number of published contents of the same type as the first content and the number of published contents of the first target to which the first content is delivered;

[0435] Step 2704: The target terminal 140 determines the first keyword based on the text description of the first content, the text description of the second content placed by the first target before the first content, and the multimodal analysis of the first content.

[0436] Step 2705: The target terminal 140 inputs the first keyword into the associated word prediction model and expands the first keyword using the knowledge graph to obtain a second keyword;

[0437] Step 2706: The target terminal 140 integrates the first keyword and the second keyword into a target keyword, and obtains the target keyword and a first weight of the target keyword;

[0438] Step 2707: The object terminal 140 sends an object tag acquisition request to the content delivery platform server;

[0439] Step 2708: the content delivery platform server 110 extracts the object tag of each object from the object tag library 150 according to the object tag request, and sends the object tag of each object in the object set to the object terminal 140;

[0440] Step 2709: The object terminal 140 determines the matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag and the second weight;

[0441] Step 2710: The object terminal 140 performs a first screening in the object set based on the matching degree between the target keyword and the object tag to obtain first screened objects, and the first screened objects are used to form a preliminary group of people to be screened;

[0442] Step 2711, the target terminal 140 obtains the first screening condition, screens the to-be-screened population based on the first screening condition, obtains the second screened objects, and forms the first screened objects into the primary delivery population;

[0443] Step 2712: The target terminal 140 obtains the attribute information of the preliminarily selected delivery population, and generates a distribution map of the attribute information of the delivery population;

[0444] Step 2713: the object terminal 140 displays the attribute information distribution diagram of the delivery population to the first object;

[0445] Step 2714: The first object A sends the second screening condition selected based on the distribution graph to the object terminal 140;

[0446] Step 2715: the target terminal 140 performs a third screening on the preliminarily selected delivery population based on the second screening condition to obtain target objects, and the target objects are combined into a final delivery population;

[0447] Step 2716: the object terminal 140 displays the final delivery group to the first object A;

[0448] Step 2717: the first object A sends a content delivery confirmation instruction to the object terminal 140;

[0449] Step 2718: the target terminal 140 delivers the first content to the target object on the content delivery platform server 110;

[0450] Step 2719, the content delivery platform server 110 sends the interest expression of the target object to the object terminal 140;

[0451] Step 2720: The target terminal 140 updates the second weight of the target object's object tag and the first weight of the target keyword of the first content based on the interest expression of the target object.

[0452] Step 2721, the object terminal 140 obtains attribute information of the target object, and determines the ratio of interest representation corresponding to each attribute information according to the attribute information;

[0453] Step 2722: The object terminal 140 displays the ratio corresponding to the attribute information to the first object A.

[0454] It can be understood that the specific process of steps 2701-2703 is similar to step 310 in the above embodiment. The specific process of steps 2704-2706 is similar to step 320 in the above embodiment. The specific process of steps 2707-2708 is similar to step 330 in the above embodiment. The specific process of steps 2709-2716 is similar to step 340 in the above embodiment. The specific process of steps 2717-2718 is similar to step 350 in the above embodiment. The specific process of step 2719 is similar to step 2210 in the above embodiment. The specific process of step 2720 is similar to step 2220 and step 2320 in the above embodiment. The specific process of steps 2721-2722 is similar to steps 2420-2440 in the above embodiment. In order to save space, it will not be repeated here.

[0455] Description of the apparatus and device of the present disclosure

[0456] It is to be understood that, although the steps in the above-mentioned flowcharts are sequentially displayed according to the characterization of arrows, these steps are not necessarily executed in sequence according to the order of arrow characterization. Unless there is a clear description in the present embodiment, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the above-mentioned flowcharts can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0457] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to the characteristics of the target object such as the target object attribute information or attribute information set, the permission or consent of the target object will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards. In addition, when the embodiment of the present application needs to obtain the attribute information of the target object, the separate permission or separate consent of the target object will be obtained through a pop-up window or jump to a confirmation page. After clearly obtaining the separate permission or separate consent of the target object, the necessary target object-related data used to enable the normal operation of the embodiment of the present application is obtained.

[0458] Fig.28 The content delivery device 2800 provided in the embodiment of the present disclosure is a schematic diagram of the structure of the content delivery device 2800. The content delivery device 2800 includes:

[0459] The condition determination unit 2810 is used to determine that the first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to deliver the first content;

[0460] A keyword determination unit 2820 is configured to, in response to the first condition being met, determine a first keyword based on the first content, expand a second keyword based on the first keyword, and integrate the first keyword and the second keyword into a target keyword;

[0461] The label acquisition unit 2830 is used to acquire the object label of each object in the object set on the content delivery platform;

[0462] An object selection unit 2840, configured to select a target object from the object set based on a matching degree between a target keyword and an object tag;

[0463] The content delivery unit 2850 is configured to deliver the first content based on the selected target object.

[0464] Optionally, the condition determination unit 2810 is specifically configured to:

[0465] If the number of published contents of the first object for delivering the first content is less than the first threshold, determining that the first condition is met;

[0466] determining a first type of first content;

[0467] Determine the number of published content of the first type on the content delivery platform;

[0468] If the number of published contents of the first type is less than the second threshold, it is determined that the first condition is met.

[0469] Optionally, the keyword determination unit 2820 is specifically configured to:

[0470] Extracting the first sub-keyword from the text description of the first content;

[0471] Extracting second sub-keywords from the text description of the second content placed by the first object before the first content;

[0472] Based on the multimodal analysis of the first content, extract the third sub-keywords;

[0473] Integrate the first sub-keyword, the second sub-keyword, and the third sub-keyword into the first keyword.

[0474] Optionally, the keyword determination unit 2820 is specifically configured to:

[0475] Input the first keyword into the associated word prediction model to obtain the fourth sub-keyword;

[0476] Use the knowledge graph to expand the first keyword to obtain the fifth sub-keyword;

[0477] Integrate the fourth and fifth sub-keywords into the second keyword.

[0478] Optionally, the target keyword has a first weight, the object tag has a second weight, and the object selection unit 2840 includes:

[0479] a matching degree determining unit (not shown), configured to determine a matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag, and the second weight;

[0480] The target object selection unit (not shown) is used to select a target object from the object set based on the matching degree of each object in the object set.

[0481] Optionally, the second weight includes a first sub-weight from the first source and a second sub-weight from the second source, and the matching degree determination unit (not shown) is specifically used to:

[0482] Determining a first matching degree corresponding to the first source based on the target keyword, the first weight, the object tag, and the first sub-weight;

[0483] Determining a second matching degree corresponding to a second source based on the target keyword, the first weight, the object tag, and the second sub-weight;

[0484] Based on the first degree of match and the second degree of match, a degree of match is determined.

[0485] Optionally, the first weight of the first keyword is a fixed weight, and the content delivery device further includes a weight determination unit (not shown), which is specifically configured to:

[0486] Determine the first weight of the second keyword, wherein the first weight of the fourth sub-keyword is the confidence probability of the fourth sub-keyword output by the associated word prediction model; and the first weight of the fifth sub-keyword is determined based on the distance between the fifth sub-keyword and the first keyword in the knowledge graph.

[0487] Optionally, the matching degree determination unit (not shown) is specifically used for:

[0488] For each object in the object set, determining a target object tag in the object tag of the object that matches one of the target keywords;

[0489] For each target object tag, determine the product of the second weight of the target object tag and the first weight of the matched target keyword;

[0490] The products of each target object label are accumulated to get the matching degree.

[0491] Optionally, the target object selection unit (not shown) is specifically used to:

[0492] In the object set, determining candidate objects whose matching degree is greater than a third threshold;

[0493] Sort the candidates in descending order of matching degree;

[0494] The candidate object ranked first in the ranking is taken as the target object.

[0495] Optionally, the label acquisition unit 2830 is specifically configured to:

[0496] For each object in the object set, obtaining a set of contents of interest represented by the object;

[0497] Obtaining the third keyword of the content in the content collection;

[0498] For each third keyword, determining the number of contents in the content set having the third keyword;

[0499] If the number of contents is greater than a fourth threshold, the third keyword is used as an object tag of the object.

[0500] Optionally, the object selection unit 2840 includes:

[0501] A first screening unit (not shown), configured to perform a first screening in the object set based on the matching degree between the target keyword and the object label to obtain a first screened object;

[0502] A screening condition acquisition unit (not shown), configured to acquire a first screening condition for a first object;

[0503] The second screening unit (not shown) is used to perform a second screening on the first screened object based on the matching of the first screening condition with the attribute information of the first screened object to obtain the target object.

[0504] Optionally, the second screening unit (not shown) is specifically used for:

[0505] Based on the matching of the first screening condition and the attribute information of the first screened object, a second screening is performed in the first screened object to obtain a second screened object;

[0506] Obtaining attribute information of the second filtered object;

[0507] Displaying a distribution diagram of objects based on attribute information after the second screening;

[0508] Displaying an attribute information checkbox control to receive a second filter condition input by the first object based on the distribution graph;

[0509] Based on the matching of the second screening condition and the attribute information of the second screened object, a third screening is performed on the second screened object to obtain the target object.

[0510] Optionally, the second screening unit (not shown) is specifically used for:

[0511] Based on the matching of the screening condition and the attribute information of the first screened object, a second screening is performed in the first screened object to obtain a second screened object;

[0512] Sort the objects after the second screening from high to low according to the matching degree;

[0513] The second screened object in the second position in the ranking is used as the first display object, and the first display object and the object label of the first display object matching the target keyword are displayed;

[0514] Selecting some of the second screened objects from the second screened objects that are not ranked in the second position in the sorting as second display objects, and displaying the second display objects and object labels of the second display objects that match the target keyword;

[0515] receiving a reserved number input by the first object, wherein the reserved number is selected by the first object according to the first display object, the object tag of the first display object matching the target keyword, the second display object, and the object tag of the second display object matching the target keyword;

[0516] The second filtered objects are retained according to the retention number to obtain the target objects.

[0517] Optionally, the content delivery device further includes a first updating unit (not shown), and the first updating unit (not shown) is used to:

[0518] Obtaining the interest expression of the target object after the first content is delivered;

[0519] According to the interest expression of the target object, the first weight of the target keyword of the first content is updated, wherein the first weight is a ratio of a first number to a second number, the first number is the number of times the target object expresses interest in the target keyword after the first content with the target keyword is delivered to the target object, and the second number is the number of times the first content with the target keyword is delivered to the target object.

[0520] Optionally, the content delivery device further includes a second updating unit (not shown), and the second updating unit (not shown) is used to:

[0521] Obtaining the interest expression of the target object after the first content is delivered;

[0522] Based on the interest expression of the target object, a second weight of the object tag of the target object is updated, wherein the second weight is a ratio of a third number to a fourth number, the third number is, with respect to the object tag, the number of times the target object expresses interest after the first content is delivered to the target object with the object tag, and the fourth number is, with respect to the object tag, the number of times the first content is delivered to the target object with the object tag.

[0523] Optionally, the content delivery device further includes a display unit (not shown), and the display unit (not shown) is used to:

[0524] Obtaining the interest expression of the target object after the first content is delivered;

[0525] Get the attribute information of the target object;

[0526] For each piece of attribute information, determining a ratio of interest in the target object corresponding to the attribute information;

[0527] The ratios corresponding to the various attribute information are displayed, and are used for screening the delivery of the third content of the first object after the delivery of the first content.

[0528] Reference Fig.29 , Fig.29 The structural block diagram of the terminal for implementing the content delivery method of the embodiment of the present disclosure is as follows: the terminal includes: a radio frequency (RF) circuit 2910, a memory 2915, an input unit 2930, a display unit 2940, a sensor 2950, ​​an audio circuit 3030, a wireless fidelity (WiFi) module 2970, a processor 2980, and a power supply 2990. Those skilled in the art can understand that Fig.29 The terminal structure shown does not constitute a limitation on the mobile phone or computer, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0529] RF circuit 2910 can be used for receiving and sending signals during information transmission or calls. In particular, after receiving the downlink information from the base station, it is sent to processor 2980 for processing; in addition, the designed uplink data is sent to the base station.

[0530] The memory 2915 may be used to store software programs and modules. The processor 2980 executes various functional applications and data processing of the target terminal by running the software programs and modules stored in the memory 2915 .

[0531] The input unit 2930 may be used to receive input digital or character information and generate key signal input related to the setting and function control of the target terminal. Specifically, the input unit 2930 may include a touch panel 2928 and other input devices 2929.

[0532] The display unit 2940 may be used to display input information or provided information and various menus of the target terminal. The display unit 2940 may include a display panel 2941.

[0533] The audio circuit 3030, the speaker 3028, and the microphone 3029 may provide an audio interface.

[0534] In this embodiment, the processor 2980 included in the terminal can execute the content delivery method of the previous embodiment.

[0535] The terminals of the embodiments of the present disclosure include but are not limited to mobile phones, computers, intelligent voice interaction devices, smart home appliances, vehicle terminals, aircraft, etc. The embodiments of the present invention can be applied to various scenarios, including but not limited to data security, blockchain, data storage, information technology, etc.

[0536] Fig.30A block diagram of the structure of part of the server for implementing the content delivery method of the embodiment of the present disclosure. The server may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 3022 (for example, one or more processors) and memory 3029, one or more storage media 2130 (for example, one or more mass storage devices) storing application programs 3042 or data 3044. Among them, the memory 3029 and the storage medium 3030 can be short-term storage or permanent storage. The program stored in the storage medium 3030 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processing unit 3022 can be configured to communicate with the storage medium 3030 and execute a series of instruction operations in the storage medium 3030 on the server.

[0537] The server may also include one or more power supplies 3030, one or more wired or wireless network interfaces 3050, one or more input and output interfaces 3058, and / or one or more operating systems 3041, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0538] The central processor 3022 in the server can be used to execute the content delivery method of the embodiment of the present disclosure.

[0539] The embodiments of the present disclosure also provide a computer-readable storage medium, which is used to store program codes, and the program codes are used to execute the content delivery methods of the aforementioned embodiments.

[0540] The embodiment of the present disclosure also provides a computer program product, which includes a computer program. A processor of a computer device reads and executes the computer program, so that the computer device executes and implements the above-mentioned transaction on-chain.

[0541] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "comprises" and "comprising" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0542] It should be understood that in the present disclosure, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0543] It should be understood that in the description of the embodiments of the present disclosure, the meaning of multiple (or multiple items) is more than two, greater than, less than, exceed, etc. are understood to not include the number, and above, below, within, etc. are understood to include the number.

[0544] In the several embodiments provided in the present disclosure, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0545] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0546] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0547] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the various embodiments of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0548] It should also be understood that the various implementations provided in the embodiments of the present disclosure can be combined arbitrarily to achieve different technical effects.

[0549] The above is a specific description of the implementation methods of the present disclosure, but the present disclosure is not limited to the above implementation methods. Technical personnel familiar with the art can also make various equivalent modifications or substitutions without violating the spirit of the present disclosure. These equivalent modifications or substitutions are all included in the scope defined by the claims of the present disclosure.

Claims

1. A content delivery method, characterized in that: include: Determining that a first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object delivering the first content; In response to the first condition being met, determining a first keyword based on the first content, expanding a second keyword based on the first keyword, and integrating the first keyword and the second keyword as a target keyword; Obtain object labels for each object in an object set on a content delivery platform; Selecting a target object from the object set based on a matching degree between the target keyword and the object tag; The first content is delivered based on the selected target object.

2. The content delivery method according to claim 1, characterized in that: The determining that the first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object delivering the first content includes: If the number of published contents of the first object that delivers the first content is less than a first threshold, determining that the first condition is met; determining a first type of the first content; Determine the number of published contents of the first type on the content delivery platform; If the number of published contents of the first type is less than a second threshold, it is determined that the first condition is met.

3. The content delivery method according to claim 1, characterized in that: The determining the first keyword based on the first content includes: Extracting a first sub-keyword from the text description of the first content; extracting second sub-keywords from a text description of a second content placed by the first object before the first content; Extracting third sub-keywords based on the multimodal analysis of the first content; The first sub-keyword, the second sub-keyword, and the third sub-keyword are integrated into the first keyword.

4. The content delivery method according to claim 1, characterized in that: The step of expanding the second keyword based on the first keyword includes: Inputting the first keyword into a related word prediction model to obtain a fourth sub-keyword; The first keyword is expanded using the knowledge graph to obtain the fifth sub-keyword; The fourth sub-keyword and the fifth sub-keyword are integrated into the second keyword.

5. The content delivery method according to claim 4, characterized in that: The target keyword has a first weight, and the object tag has a second weight; The selecting a target object from the object set based on the matching degree between the target keyword and the object tag includes: Determining a matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag, and the second weight; A target object is selected from the object set based on the matching degree of each of the objects in the object set.

6. The content delivery method according to claim 5, characterized in that: The second weight includes a first sub-weight from a first source and a second sub-weight from a second source; The determining the matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag, and the second weight includes: Determining a first matching degree corresponding to the first source based on the target keyword, the first weight, the object tag, and the first sub-weight; determining a second matching degree corresponding to the second source based on the target keyword, the first weight, the object tag, and the second sub-weight; The degree of matching is determined based on the first degree of matching and the second degree of matching.

7. The content delivery method according to claim 5, characterized in that: The first weight of the first keyword is a fixed weight; After expanding the second keyword based on the first keyword, the content delivery method also includes: determining the first weight of the second keyword, wherein the first weight of the fourth sub-keyword is the confidence probability of the fourth sub-keyword output by the associated word prediction model; the first weight of the fifth sub-keyword is determined based on the distance between the fifth sub-keyword and the first keyword in the knowledge graph.

8. The content delivery method according to claim 5, characterized in that: The determining the matching degree between the target keyword and the object tag based on the target keyword, the first weight, the object tag and the second weight includes: For each of the objects in the object set, determining a target object tag in the object tags of the object that matches one of the target keywords; For each target object tag, determine a product of the second weight of the target object tag and the first weight of the matching target keyword; The products of each target object label are accumulated to obtain the matching degree.

9. The content delivery method according to claim 5, characterized in that: The selecting a target object from the object set based on the matching degree of each object in the object set includes: In the object set, determining candidate objects whose matching degree is greater than a third threshold; Sorting the candidate objects in descending order of the matching degree; The candidate object in the first position in the ranking is used as the target object.

10. The content delivery method according to claim 1, characterized in that: The obtaining of the object label of each object in the object set on the content delivery platform includes: For each of the objects in the object set, obtaining a content set of interest represented by the object; Obtaining a third keyword of the content in the content set; For each of the third keywords, determining the number of contents in the content set having the third keyword; If the number of contents is greater than a fourth threshold, the third keyword is used as the object tag of the object.

11. The content delivery method according to claim 1, characterized in that: The selecting a target object from the object set based on the matching degree between the target keyword and the object tag includes: Based on the matching degree between the target keyword and the object label, performing a first screening in the object set to obtain a first screened object; Obtaining a first screening condition for the first object; Based on the matching of the first screening condition and the attribute information of the first screened object, a second screening is performed on the first screened object to obtain the target object.

12. The content delivery method according to claim 11, characterized in that: The step of performing a second screening on the first screened object based on the matching of the first screening condition with the attribute information of the first screened object to obtain the target object includes: Based on the matching of the first screening condition and the attribute information of the first screened object, performing a second screening on the first screened object to obtain a second screened object; Obtaining attribute information of the second filtered object; Displaying a distribution diagram of the second filtered objects based on the attribute information; Displaying an attribute information checkbox to receive a second filter condition input by the first object based on the distribution graph; Based on the matching of the second screening condition and the attribute information of the second screened object, a third screening is performed on the second screened object to obtain the target object.

13. The content delivery method according to claim 11, characterized in that: The step of performing a second screening on the first screened object based on the matching of the first screening condition with the attribute information of the first screened object to obtain the target object includes: Based on the matching of the screening condition and the attribute information of the first screened object, performing a second screening in the first screened object to obtain a second screened object; sorting the second screened objects from high to low according to the matching degree; taking the second screened object in the top second position in the ranking as the first display object, and displaying the first display object and the object label of the first display object matching the target keyword; Selecting some of the second screened objects from the second screened objects that are not ranked in the second top position in the ranking as second display objects, and displaying the second display objects and the object tags of the second display objects that match the target keyword; receiving a reserved number input by the first object, wherein the reserved number is selected by the first object according to the first display object, the object tag of the first display object matching the target keyword, the second display object, and the object tag of the second display object matching the target keyword; The second screened objects are retained according to the retained number to obtain the target objects.

14. The content delivery method according to claim 1, characterized in that: After delivering the first content based on the selected target object, the content delivery method further includes: Acquire the interest expression of the target object after delivering the first content; According to the interest expression of the target object, the first weight of the target keyword of the first content is updated, wherein the first weight is a ratio of a first number to a second number, the first number is the number of times the target object expresses interest in the target keyword after the first content with the target keyword is delivered to the target object, and the second number is the number of times the first content with the target keyword is delivered to the target object.

15. The content delivery method according to claim 1, characterized in that: After delivering the first content based on the selected target object, the content delivery method further includes: Acquire the interest expression of the target object after delivering the first content; According to the interest expression of the target object, a second weight of the object tag of the target object is updated, wherein the second weight is a ratio of a third number to a fourth number, the third number is the number of times the target object expresses interest after the first content is delivered to the target object with the object tag, with respect to the object tag, and the fourth number is the number of times the first content is delivered to the target object with the object tag, with respect to the object tag.

16. The content delivery method according to claim 1, characterized in that: After delivering the first content based on the selected target object, the content delivery method further includes: Acquire the interest expression of the target object after delivering the first content; Acquire attribute information of the target object; For each piece of the attribute information, determining a ratio of the target object corresponding to the attribute information that indicates interest; The ratios corresponding to the respective attribute information are displayed for screening the delivery of the third content of the first object after the delivery of the first content.

17. A content delivery device, characterized in that: include: A condition determination unit, configured to determine that a first condition is satisfied based on the number of published contents of the same type as the first content to be delivered and the number of published contents of the first object to which the first content is delivered; a keyword determination unit, configured to, in response to the first condition being met, determine a first keyword based on the first content, expand a second keyword based on the first keyword, and integrate the first keyword and the second keyword as a target keyword; A label acquisition unit, used to acquire an object label of each object in an object set on a content delivery platform; an object selection unit, configured to select a target object from the object set based on a matching degree between the target keyword and the object tag; A content delivery unit is configured to deliver the first content based on the selected target object.

18. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the content delivery method according to any one of claims 1 to 16 is implemented.

19. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the content delivery method according to any one of claims 1 to 16 is implemented.

20. A computer program product, comprising a computer program, wherein the computer program is read and executed by a processor of a computer device, so that the computer device executes the content delivery method according to any one of claims 1 to 16.

Citation Information

Patent Citations

  • Goods information and keyword matching method, and goods information releasing method and device

    CN103778548A

  • Information push method, apparatus and device, and storage medium

    CN106982256A

  • An advertisement putting method and platform

    CN109919641A

  • Advertisement putting method, device and equipment

    CN110020880A

  • Target user screening method and device

    CN110717788A