Advertisement delivery method, device, electronic device and storage medium

Abstract

The present application relates to an advertising delivery method, device, electronic device and storage medium, the method comprising: receiving an advertisement to be delivered, determining the advertising element information corresponding to the advertisement, the advertising element information including a plurality of element entities; obtaining graph data from a preset graph database, the graph data being formed by entity extraction and knowledge fusion of corresponding user portrait data; determining the target user based on the relevance of the plurality of element entities to all entities associated with the target entity in the graph data, the target entity being a named entity used to characterize the user name; determining a delivery strategy based on the user behavior information of the target user, and delivering the advertisement to the target user based on the delivery strategy. Through the present application, the problems of poor delivery effect and high delivery cost of advertisements delivered in a large-scale, fixed-time delivery method in the related art are solved, and the beneficial effect of reducing the advertising costs of advertisers and improving the delivery effect of advertisements is achieved through precise delivery.

Description

Technical Field

[0001] The present application relates to the field of Internet advertising technology, and in particular to an advertisement delivery method, device, electronic device, and storage medium. Background Art

[0002] With the rapid development of the internet and mobile internet, the web has become an indispensable part of people's lives, allowing them to access a wide variety of information. As a crucial information medium, the internet hosts a massive amount of advertising every day. Various types of ads are delivered through various internet advertising platforms, such as pop-up and floating ads on web pages, pre-roll ads at the beginning of videos, and mid-roll ads.

[0003] In the prior art, internet advertising is typically delivered over a large scale and at fixed times. For example, a specific ad might be placed before a video on a video website within a week. This large-scale, fixed-time approach fails to deliver targeted advertising to specific audiences within a specific, smaller area, reducing advertising effectiveness and increasing costs.

[0004] There is no effective solution to the problem that related technologies use a large-scale, fixed-time delivery method to deliver advertisements, resulting in poor delivery effects and high delivery costs. Summary of the Invention

[0005] The present application provides an advertisement delivery method, device, electronic device and storage medium to at least solve the problem in the related art of poor delivery effect and high delivery cost of advertisements delivered in a large-scale, fixed-time delivery manner.

[0006] In the first aspect, the present application provides an advertising delivery method, comprising: receiving an advertisement to be delivered, determining the advertising element information corresponding to the advertisement, wherein the advertising element information includes multiple element entities; obtaining graph data from a preset graph database, wherein the graph data is formed by entity extraction and knowledge fusion of corresponding user portrait data; determining a target user based on the correlation between the multiple element entities and all entities associated with the target entity in the graph data, wherein the target entity is a named entity used to represent the user name; determining a delivery strategy based on the user behavior information of the target user, and delivering the advertisement to the target user based on the delivery strategy.

[0007] In a second aspect, the present application provides an intention recognition device, comprising:

[0008] A receiving module, configured to receive an advertisement to be delivered and determine advertisement element information corresponding to the advertisement, wherein the advertisement element information includes a plurality of element entities;

[0009] An acquisition module is used to obtain graph data from a preset graph database, wherein the graph data is formed by performing entity extraction and knowledge fusion on the corresponding user portrait data;

[0010] a determination module, configured to determine a target user based on the relevance of the plurality of element entities and all entities associated with a target entity in the graph data, wherein the target entity is a named entity used to represent a user name;

[0011] The processing module is used to determine a delivery strategy according to the user behavior information of the target user, and deliver the advertisement to the target user based on the delivery strategy.

[0012] In a third aspect, an electronic device is provided, comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0013] Memory for storing computer programs;

[0014] The processor is configured to implement the steps of the advertising delivery method described in any one of the embodiments of the first aspect when executing the program stored in the memory.

[0015] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the advertising delivery method as described in any embodiment of the first aspect are implemented.

[0016] This application can be applied to the field of Internet advertising for targeted identification of delivery targets and precise delivery of advertisements. The advertising delivery method, device, electronic device and storage medium provided in the embodiment of this application, by receiving the advertisement to be delivered, determine the advertising element information corresponding to the advertisement, and the advertising element information includes multiple element entities; obtain graph data from a preset graph database, and the graph data is formed by entity extraction and knowledge fusion of the corresponding user portrait data; determine the target user according to the relevance of all entities associated with the target entity in the graph data, and the target entity is a named entity used to characterize the user name; determine the delivery strategy according to the user behavior information of the target user, and deliver the advertisement to the target user based on the delivery strategy, solving the problem of poor delivery effect and high delivery cost of advertising delivered in a large-scale, fixed-time delivery method in the related technology, and achieving the beneficial effect of reducing the advertising costs of advertisers and improving the delivery effect through precise delivery.

[0017] The details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0020] Figure 1 A flowchart of an advertisement delivery method provided in an embodiment of the present application;

[0021] Figure 2 Schematic diagram of NLP entity extraction in a preferred embodiment of the present application;

[0022] Figure 3 This is a schematic diagram of the atlas data of the preferred embodiment of the present application Figure 1 ;

[0023] Figure 4 This is a schematic diagram of the atlas data of the preferred embodiment of the present application Figure 2 ;

[0024] Figure 5 It is a structural diagram of an advertising delivery device according to an embodiment of the present application;

[0025] Figure 6 Schematic diagram of the structure of the electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0027] Before describing the embodiments of the present application, the relevant technical means used in the advertising delivery method of the embodiments of the present application and the problems existing in the relevant technologies are explained as follows.

[0028] Natural Language Processing (NLP) is a technology that uses computers to process and manipulate information in written and spoken natural languages, which are unique to humans. NLP is used in machine translation, automated question answering, speech processing, information retrieval, and information extraction.

[0029] Knowledge extraction refers to extracting the knowledge contained in the information source through processes such as identification, understanding, screening, and induction, and storing it to form a knowledge meta-repository; that is, extracting knowledge from data of different sources and structures, forming knowledge and storing it in the knowledge graph. The knowledge graph includes entities and weights. The weight is the relationship coefficient between entities, which can be calculated through different keyword algorithms in NLP technology (for example: TF-IDF, Text-Rank).

[0030] A knowledge graph is a semantic network that describes objective things in the form of a graph, composed of nodes and edges. Nodes in a knowledge graph represent concepts and entities, with concepts being abstract and entities being concrete. Edges represent relationships and attributes between things. Attributes represent internal characteristics of things, while relationships represent external connections. A knowledge graph stores data in triplets of "entity-relationship-entity" or "entity-attribute-value," forming a graphical knowledge base.

[0031] The various techniques described in this application can be used for named entity recognition, keyword extraction, information extraction, etc.

[0032] Figure 1 This is a flow chart of an advertisement delivery method provided in an embodiment of the present application. Figure 1 As shown, an embodiment of the present application provides an advertisement delivery method, which includes the following steps:

[0033] Step S101: receiving an advertisement to be delivered, and determining advertisement element information corresponding to the advertisement, wherein the advertisement element information includes a plurality of element entities.

[0034] In this embodiment, the subject that executes the advertising delivery method of this application is an advertising delivery platform, which can also be an advertising delivery terminal; in this embodiment, after the advertisement provider (advertiser) provides the advertisement to the corresponding delivery end (delivery platform, terminal), the delivery end will first process the data of the advertisement content to obtain advertisement element information including multiple element entities; in this embodiment, the advertisement includes at least advertisement attributes and advertisement elements, wherein the advertisement attributes include type, applicable group (age), etc.; the advertisement elements include spokesperson, brand, product category, etc.

[0035] In some optional implementations, the means of obtaining advertising element information can be manual extraction, that is, when providing an advertisement, corresponding advertising elements describing the advertisement are also provided, such as: keywords of the advertisement; NLP technology can also be used to obtain advertising element information and element entities; for example: the advertisement is: star A endorses product XX, and the obtainable element entities are star A and product XX.

[0036] Step S102: Obtain graph data from a preset graph database, wherein the graph data is formed by performing entity extraction and knowledge fusion on the corresponding user portrait data.

[0037] In this embodiment, user portrait data is obtained by integrating the user's static information data and dynamic information data, wherein the user's static information data includes data on demographic attributes, business attributes, such as age, gender, income, occupation, and region, and the static information data can itself become a user label; the user's dynamic information data is the user's constantly changing behavior information. When the user behavior is presented on the Internet, it is manifested as browsing web pages or product pages, searching for products, posting Weibo and Weibo messages about products, etc.; in this embodiment, after collecting the user's static information data and dynamic information data, data modeling is performed to construct a user portrait data model, and the following formula can be used to construct the user portrait data: user identification + time + behavior type + contact point (URL + content), that is, to describe a user because of what time, place, and what he did, so he will be labeled with MM.

[0038] In this embodiment, NLP is used to extract entities from dynamic information data in user portrait data. Figure 2 This is a schematic diagram of NLP entity extraction in the preferred embodiment of the present application, refer to Figure 2 In some optional embodiments, entity extraction is performed by collecting the user's historical search records, historical browsing records, current search content, etc. Entity extraction is also known as named entity recognition, including entity detection and classification, such as identifying names of people and places.

[0039] In some optional embodiments of this embodiment, it is also possible to collect user search records (for example: "Star B TV series"), browsed video information (such as: "Character A turned to look at Character B, and Character C's reaction was so real"), advertising information (such as "Star C endorses cosmetics brand A"), and live broadcast information ("Character C cosmetics brand B live fan shot video", "Cosmetics brand C-Star D raids the dressing room"), and use the above data as analysis objects to perform entity extraction, and perform the following steps: segment the text corpus to be extracted; obtain the field labels to be identified, and label the segmentation results; extract the labeled segmented words; and organize the extracted segmented words into named entities of the required field; in this way, the entities that can be extracted are: character A, character B, character C, cosmetics brand A, brand, and dressing room.

[0040] In this embodiment, the extraction of entities is completed using NLP technology. Through NLP's precise analysis of the search content, and based on the knowledge elements such as entities, relationships, and attributes extracted from the relevant data (including structured data and semi-structured data) obtained from the corresponding public platform, a knowledge graph is constructed. In this way, the user's preference direction can be obtained more accurately. When the advertisement to be delivered matches the corresponding user's preference, the advertisement is delivered to the user at this time, which can achieve accurate delivery of advertisements and improve the effectiveness of advertisement delivery.

[0041] In this embodiment, in order to improve the efficiency of accurate advertising delivery, the graph data obtained from the preset graph database (reference Figure 3 and Figure 4 The graph data shown) is graph data that is preliminarily screened out from multiple graph data in a graph database based on multiple element entities, that is, the selected graph data is graph data that has a certain relevance to the advertisement to be placed. For example: the advertisement to be placed is about the endorsement of a certain celebrity, and the corresponding graph data is that a user has searched for the celebrity in the historical Internet usage process, and the corresponding user's search record contains film and television products, advertising information, live broadcast information, and Weibo messages related to the celebrity; by preliminarily screening the graph data, the data can be more matched when calculating the relationship between multiple element entities and users, and the target users can be determined more accurately.

[0042] Step S103: determining a target user based on the relevance of the multiple element entities and all entities associated with the target entity in the graph data, wherein the target entity is a named entity used to represent the user name.

[0043] In this embodiment, after obtaining the graph data, the relationship between the feature entity and the user corresponding to the graph data is calculated, and based on the degree of correlation between the feature entity and the corresponding user, it is determined whether the user's preference matches the advertisement to be delivered. For example: by matching the feature entity with the corresponding graph data, users with a high degree of match are screened out, indicating that the user has a strong willingness or preference to receive the delivered advertisement. At this time, the delivered advertisement is likely to be responded to by the user, thereby improving the efficiency of advertising delivery.

[0044] In this embodiment, the relationship between each element entity and the user is calculated to determine whether the correlation between the advertisement to be delivered and the user is high. The relationship between the element entity and the user is calculated by multiplying the relationship coefficients between each element entity and the entities associated with the user. It should be noted that when calculating the correlation between the element entity and the entity associated with the user, it is necessary to first calculate the product of the relationship coefficients of the element entity and the named entity in the entity corresponding to the user, and then multiply it by the relationship coefficient between the named entity representing the user name and the corresponding entity to obtain the relationship coefficient between the element entity and the corresponding entity. If the corresponding entity is also associated with the corresponding entity at the next level, it is necessary to calculate the relationship coefficient between the element entity and the corresponding entity at the next level. In this way, The need to calculate the relationship between each element entity and the user; for example: when the entity representing the user name is entity A, the entity A is associated with entity B to entity M, and there is at least one associated entity a at the next level from entity B to entity M, that is, the corresponding entity a is connected to entity A through one of the corresponding entities from entity B to entity M; for example: in the process of calculating the relationship between each element entity and the user, when calculating the relationship coefficient between the element entity and entity a, it is necessary to calculate the relationship coefficient between entity A and the element entity, the relationship coefficient between entity A and the corresponding one of the associated entities from B to entity M, and the relationship coefficient between the corresponding one of entity B to entity M and entity a. The product of all the relationship coefficients corresponds to the relationship coefficient between the element entity and entity a.

[0045] In this embodiment, when the target user matching degree corresponding to the corresponding graph data calculated based on the element entity is high, the confirmation of the corresponding delivery target is completed.

[0046] Step S104: Determine a delivery strategy based on the user behavior information of the target user, and deliver the advertisement to the target user based on the delivery strategy.

[0047] In this embodiment, after determining the corresponding user target, a delivery strategy is further generated based on the corresponding user's operations. The corresponding delivery strategy includes determining the delivery object and / or delivery time of the advertisement based on the corresponding user's historical habits of delivery device / platform and user time, thereby achieving more accurate advertisement delivery.

[0048] Through the above steps S101 to S104, the advertisement to be delivered is received, and the advertising element information corresponding to the advertisement is determined, and the advertising element information includes multiple element entities; the graph data is obtained from the preset graph database, and the graph data is formed by entity extraction and knowledge fusion of the corresponding user portrait data; the target user is determined according to the relevance of the multiple element entities and all entities associated with the target entity in the graph data, and the target entity is a named entity used to represent the user name; the delivery strategy is determined according to the user behavior information of the target user, and the advertisement is delivered to the target user based on the delivery strategy, which solves the problems of poor delivery effect and high delivery cost of advertising delivered in a large-scale, fixed-time delivery method in related technologies, and achieves the beneficial effect of reducing advertisers' advertising costs and improving advertising delivery effects through precise delivery.

[0049] In some embodiments, determining the target user based on the relevance of the multiple element entities to all entities associated with the target entity in the graph data in step S103 can be achieved by the following steps:

[0050] Step 21: Extract the target entity corresponding to the graph data and all entities associated with the target entity.

[0051] In this embodiment, the graph data is at least displayed according to the relationship and attributes between entities, and the corresponding entities can be directly extracted from the graph data.

[0052] Step 22: Calculate the relationship coefficient between each feature entity and the target entity and all entities associated with the target entity in turn, and determine the sum of the relationship coefficients between each feature entity and the target entity and all entities associated with the target entity, where the relationship coefficient is used to characterize the degree of association between the corresponding entity and the feature entity.

[0053] In this embodiment, NLP technology is used to calculate the relationship coefficients between the feature entity and the target entity and entities associated with the target entity, that is, to determine the corresponding word weights or relationship weights.

[0054] In this embodiment, when calculating the relationship coefficient between the element entity and all entities, when the current entity is also associated with the next-level entity, it is necessary to determine the corresponding relationship coefficient based on the relationship coefficient of the current entity element entity and the relationship coefficient of the current element entity with the associated next-level entity. For example: when the target entity is entity 1, the entity 1 is associated with entity 2 to entity N, and one of entity 2 to entity N is associated with the next-level entity n, so that entity n is connected to entity 1 through one of entity 2 to entity N; for example: in the process of calculating the relationship coefficient between each element entity and all entities, when calculating the relationship weight between the element entity and entity n, it is necessary to calculate the relationship coefficient between entity 1 and the element entity, the relationship coefficient between entity 1 and the corresponding one of the associated entities 2 to entity N, and the relationship coefficient between the corresponding one of entity 2 to entity N and entity n. The product of all the relationship coefficients corresponds to the relationship coefficient between the element entity and entity n.

[0055] In this embodiment, when calculating the relationship between the element entity and the user corresponding to the target entity, that is, determining the correlation between the element entity and the user, it is necessary to sum the relationship coefficients of all entities associated with the element entity and the target entity. When the sum of the relationship coefficients is greater than the set threshold, it indicates that the user has a high degree of matching with the element entity and the corresponding advertisement.

[0056] Step 23: Determine whether the sum of the relationship coefficients is greater than a preset threshold.

[0057] Step 24: When it is determined that the sum of the relationship coefficients is greater than a preset threshold, the user corresponding to the target entity is determined to be the target user.

[0058] The target entity corresponding to the graph data extracted in the above steps and all entities associated with the target entity are calculated in turn; the relationship coefficients of each element entity with the target entity and all entities associated with the target entity are calculated, and the sum of the relationship coefficients of each element entity with the target entity and all entities associated with the target entity is determined; it is judged whether the sum of the relationship coefficients is greater than a preset threshold; when it is judged that the sum of the relationship coefficients is greater than the preset threshold, the user corresponding to the target entity is determined to be the target user, and the relevance of the element entity corresponding to the advertising content and the graph data corresponding to the target user is matched, so that the advertisement to be delivered is the content preferred by the target user, thereby accurately delivering the advertisement and improving the effective delivery rate of the advertisement.

[0059] In some embodiments, the step 21 of sequentially calculating the relationship coefficients between each element entity and the target entity and all entities associated with the target entity may be implemented by the following steps:

[0060] Step 31: Determine a first entity among all entities associated with the target entity and a second entity associated with the first entity, wherein the first entity is an entity directly associated with the target entity.

[0061] In this embodiment, the first entity is an entity that has a direct relationship with the target entity, that is, in the graph data, the target entity and the first entity are respectively on two directly connected nodes, and the second entity can be an entity that has a direct relationship with the first entity, that is, the first entity and the second entity are on two directly connected nodes, or it can be an indirectly connected entity, that is, there is an intermediate entity between the first entity and the second entity that associates the two entities. Of course, the second entity is at the boundary of the graph data. At the same time, the first entity may not be associated with the corresponding second entity. At this time, the first entity is at the boundary of the graph data; in some optional implementation schemes of the present application, the graph data is hierarchical in the following format: user's corresponding target entity-first entity-second entity, that is, no corresponding intermediate entity is set between the second entity and the first entity, and the second entity is also the corresponding boundary of the graph data.

[0062] Step 32: Calculate the first relationship coefficient between the element entity and the target entity, the second relationship coefficient between the target entity and the first entity, and the third relationship coefficient between the first entity and the second entity by using natural language processing (NLP).

[0063] In this embodiment, the use of natural language processing NLP to calculate the corresponding relationship coefficient should be understood as a known means, and the existing algorithms and means used to determine the corresponding relationship coefficient are applicable to this application.

[0064] Step 33: Determine the relationship coefficients between the feature entity, the target entity, and all entities associated with the target entity based on the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient.

[0065] In this embodiment, the first relationship coefficient corresponds to the relationship coefficient between the target entity and the element entity. The relationship coefficient between the element entity and the first entity is determined based on the first relationship coefficient and the second relationship coefficient, and the relationship coefficient between the element entity and the second entity is determined based on the first relationship coefficient, the second relationship coefficient and the third relationship coefficient; when there is an intermediate-level entity between the first entity and the second entity, the corresponding third relationship coefficient is the product of the relationship coefficient between the first entity and the intermediate-level entity and the relationship coefficient between the intermediate-level entity and the second entity.

[0066] Through the above steps, the first entity among all entities associated with the target entity and the second entity associated with the first entity are determined; natural language processing NLP is used to calculate the first relationship coefficient between the feature entity and the target entity, the second relationship coefficient between the target entity and the first entity, and the third relationship coefficient between the first entity and the second entity; based on the first relationship coefficient, the second relationship coefficient and the third relationship coefficient, the relationship coefficient between the feature entity and the target entity and all entities associated with the target entity is determined, the calculation of the relationship coefficient between the feature entity and the entity associated with the corresponding user is realized, and the product of the relationship coefficients is used to determine the relationship coefficient between two entities associated through a common entity, thereby ensuring the effectiveness of the determination of the correlation between entities, so that when the advertising elements are matched with the graph data, the matching degree of the matched users is high.

[0067] In some embodiments, determining the relationship coefficients of the feature entity with the target entity and all entities associated with the target entity based on the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient in step 33 is achieved by the following steps:

[0068] Step 41: Determine the first relationship coefficient as the relationship coefficient corresponding to the feature entity and the target entity.

[0069] Step 42: multiply the first relationship coefficient and the second relationship coefficient to obtain a second coefficient, wherein the relationship coefficient corresponding to the element entity and the first entity includes the second coefficient;

[0070] Step 43: Multiply the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient to obtain a third coefficient, and determine that the relationship coefficient corresponding to the element entity and the second entity includes the third coefficient.

[0071] Through the above steps, the first relationship coefficient is determined as the relationship coefficient corresponding to the feature entity and the target entity; the first relationship coefficient and the second relationship coefficient are multiplied, and the second coefficient obtained is used as the relationship coefficient corresponding to the feature entity and the first entity; the first relationship coefficient, the second relationship coefficient and the third relationship coefficient are multiplied, and the third coefficient obtained is used as the relationship coefficient corresponding to the feature entity and the second entity, thereby realizing the assignment of the relationship coefficients between the corresponding entities.

[0072] In some embodiments, based on determining the relationship coefficients between corresponding entities, determining the sum of the relationship coefficients between each element entity and the target entity and all entities associated with the target entity in step 22 can be achieved by the following steps:

[0073] Step 51: Obtain a first relationship coefficient, a second coefficient, and a third coefficient.

[0074] Step 52: Accumulate the first relationship coefficient, the second coefficient, and the third coefficient to obtain a total relationship weight, wherein the sum of the relationship coefficients between each element entity and the target entity and all entities associated with the target entity includes the total relationship weight.

[0075] In this embodiment, after the relationship weight sum is accumulated, when the relationship weight sum is greater than a preset threshold, it is determined that the user corresponding to the target entity has a high relevance and is locked as the target user.

[0076] Through the above steps 51 to 52, the relevance between the element entity and the user is determined, so as to facilitate the subsequent locking of the target user and achieve accurate advertising delivery.

[0077] In some embodiments, determining the delivery strategy based on the user behavior information of the target user in step S104 is implemented by the following steps:

[0078] Step 61: Determine the action tag information corresponding to the user based on the acquired user behavior information, wherein the action tag information includes preference information for characterizing the user's historical operations.

[0079] In this embodiment, after the target user is determined, precise delivery will be carried out according to the user's corresponding Internet usage habits. The user's corresponding Internet usage habits include the devices and platforms commonly used by the user, such as: commonly used media platforms, the user's habit of logging in on the PC or mobile terminal, and the habitual Internet usage time period, such as: the habitual online or Internet time; when the habits corresponding to the target user are analyzed and obtained, the advertisement is delivered while catering to the user's habits, so as to improve the effective delivery of the advertisement, that is, the delivered advertisement is preferred by the user, so that the corresponding user can pay attention to and / or respond to the delivered advertisement.

[0080] In this embodiment, action label information is used in user behavior information to represent the user's historical operation preferences, that is, the above-mentioned network usage habits. In addition, the action label information can also be stored in a preset graph database as the user's corresponding graph data, and retained in the form of graph data composed of multiple named entities, thereby ensuring the consistency of the database system data.

[0081] Step 62: Based on the preference information of the user's historical operations, determine the classification information corresponding to the user's historical operations, wherein the classification information corresponding to the user's historical operations includes at least one of the following: operation device object information and operation time information.

[0082] In this embodiment, the corresponding parameters for user habits include two types: time-type parameters, and hardware and software equipment such as devices / components / platforms, specifically including: the media platforms commonly used by users and the time periods when they are accustomed to using the Internet; by determining the type of preference information of the user's historical operations obtained, and then determining the receiving end after the advertisement is delivered and the time when the advertisement is delivered, when the corresponding user receives the advertisement, if the advertisement is delivered based on his or her own Internet usage habits, this will increase the user's acceptance of the advertisement and reduce resistance and boycott of the advertisement.

[0083] Step 63: Determine a delivery strategy based on the classification information, wherein the delivery strategy includes delivery devices and / or delivery time for advertisement delivery.

[0084] According to the user behavior information obtained in the above steps, the action tag information corresponding to the user is determined, and the action tag information includes preference information for characterizing the user's historical operations. Based on the preference information of the user's historical operations, the classification information corresponding to the user's historical operations is determined; based on the classification information, the delivery strategy is determined, wherein the delivery strategy includes the delivery device and / or delivery time for advertising delivery, so as to determine the corresponding delivery object and / or delivery time based on the user's Internet usage habits, further improve the user's acceptance of advertising, reduce resistance and opposition to advertising delivery, and meet the requirements of precise delivery while reducing advertising delivery costs and improving advertising delivery effects.

[0085] In some embodiments, the process of constructing the atlas data includes the following steps:

[0086] Step 71: Obtain user portrait data corresponding to a preset user and text data corresponding to a preset scenario, wherein the user portrait data is generated based on user tag information and user behavior data, and the text data includes structured data and partially structured data.

[0087] In this embodiment, by collecting user portrait data and text data, corresponding data preparation is provided for subsequent knowledge extraction and knowledge fusion, so that the corresponding graph data matches the corresponding preset scenario.

[0088] Step 72: Identify and extract entities and entity attributes from the user portrait data and text data respectively to obtain at least two entities and attribute information of each entity.

[0089] In this embodiment, the identification and extraction of entities and entity attributes from text data is performed after entity extraction from user portrait data through NLP, that is, after obtaining text data corresponding to a preset scenario, data related to the entities extracted by NLP is detected in the text data, and then knowledge extraction is performed on the detected related data. Specifically, the following processing is included: named entity recognition, term extraction, relationship extraction, event extraction, and coreference resolution. For example, the entity "Beijing" and the corresponding classification category entity "place name" are detected from the corpus "Beijing is a busy city". For example, multiple single-component related terms are found from the corpus to complete term extraction. For example, relationship extraction is performed on the corpus "Character W and Character X are good friends", and the following can be obtained: [Character W] <Friend> [Character X]. For example, the trigger words, time, place, and other information of the event are extracted from a news report. For example, the referents of the pronouns "he", "she", and "it" in the corpus are determined, and coreference resolution is completed.

[0090] In this embodiment, knowledge is extracted from user portrait data and text data to obtain corresponding entities and the relationships and attributes between entities.

[0091] Step 73: Perform knowledge fusion processing on the relationship between any two entities and the attribute information of each entity to generate at least one entity-to-relationship template, wherein the graph data includes the entity-to-relationship template, and the knowledge fusion processing includes entity similarity calculation, deduplication, and disambiguation.

[0092] In this embodiment, generating entity pair relationship templates is also a process of knowledge extraction.

[0093] Through the above steps, user portrait data corresponding to the preset user and text data corresponding to the preset scenario are obtained; entities and entity attributes are identified and extracted from the user portrait data and text data respectively to obtain at least two entities and attribute information of each entity; knowledge fusion processing is performed on the relationship between any two entities and the attribute information of each entity to generate at least one entity-to-relationship template. The graph data includes the entity-to-relationship template. The knowledge fusion processing includes entity similarity calculation, deduplication, and disambiguation. User image data collection, NLP entity extraction, knowledge extraction and knowledge fusion are used to realize the generation of graph data. Moreover, through the precise analysis of the retrieval data content by NLP and the construction of knowledge graph data based on the text data corresponding to the preset scenario, the user's preference direction can be obtained more accurately, so that the target user can be accurately matched and the advertisement can be accurately delivered.

[0094] It should be noted that the steps shown in the above process or the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0095] This embodiment also provides an advertising delivery device for implementing the above-mentioned embodiments and preferred implementations. Details already described will not be repeated. As used below, terms such as "module," "unit," and "subunit" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0096] Figure 4 is a structural block diagram of an advertisement delivery device according to an embodiment of the present application. Figure 4 As shown, the device includes:

[0097] The receiving module 51 is configured to receive an advertisement to be delivered and determine advertisement element information corresponding to the advertisement, wherein the advertisement element information includes a plurality of element entities;

[0098] The acquisition module 52 is coupled to the receiving module 51 and is used to obtain graph data from a preset graph database, wherein the graph data is formed by performing entity extraction and knowledge fusion on the corresponding user portrait data;

[0099] a determination module 53 coupled to the acquisition module 52 and configured to determine a target user based on the relevance of the plurality of element entities to all entities associated with the target entity in the graph data, wherein the target entity is a named entity representing a user name;

[0100] The processing module 54 is coupled to the determination module 53 and is configured to determine a delivery strategy according to the user behavior information of the target user, and deliver the advertisement to the target user based on the delivery strategy.

[0101] Through the device of the embodiment of the present application, an advertisement to be delivered is received, and the advertising element information corresponding to the advertisement is determined, and the advertising element information includes multiple element entities; graph data is obtained from a preset graph database, and the graph data is formed by entity extraction and knowledge fusion of corresponding user portrait data; the target user is determined based on the correlation between the multiple element entities and all entities associated with the target entity in the graph data, and the target entity is a named entity used to represent the user name; the delivery strategy is determined based on the user behavior information of the target user, and the advertisement is delivered to the target user based on the delivery strategy, which solves the problems of poor delivery effect and high delivery cost of advertising delivered in a large-scale, fixed-time delivery method in related technologies, and achieves the beneficial effect of reducing advertisers' advertising costs and improving advertising delivery effects through precise delivery.

[0102] In some embodiments, the determination module 32 further includes:

[0103] A first extraction unit is used to extract the target entity corresponding to the graph data and all entities associated with the target entity;

[0104] a first calculation unit, coupled to the first extraction unit, configured to sequentially calculate a relationship coefficient between each element entity and the target entity and all entities associated with the target entity, and determine a sum of relationship coefficients between each element entity and the target entity and all entities associated with the target entity, wherein the relationship coefficient is used to represent a degree of association between the corresponding entity and the element entity;

[0105] a first judging unit, coupled to the first calculating unit, for judging whether the sum of the relationship coefficients is greater than a preset threshold;

[0106] The first determining unit is coupled to the first judging unit and is configured to determine that the user corresponding to the target entity is the target user when it is determined that the sum of the relationship coefficients is greater than a preset threshold.

[0107] In some embodiments, the first extraction unit is also used to determine the first entity among all entities associated with the target entity and the second entity associated with the first entity, wherein the first entity is the entity directly associated with the target entity; using natural language processing NLP, calculate the first relationship coefficient between the element entity and the target entity, the second relationship coefficient between the target entity and the first entity, and the third relationship coefficient between the first entity and the second entity; based on the first relationship coefficient, the second relationship coefficient and the third relationship coefficient, determine the relationship coefficient between the element entity and the target entity and all entities associated with the target entity.

[0108] In some embodiments, the first extraction unit is also used to determine the first relationship coefficient as the relationship coefficient corresponding to the element entity and the target entity; multiply the first relationship coefficient and the second relationship coefficient to obtain the second coefficient, wherein the relationship coefficient corresponding to the element entity and the first entity includes the second coefficient; multiply the first relationship coefficient, the second relationship coefficient and the third relationship coefficient to obtain the third coefficient, and determine that the relationship coefficient corresponding to the element entity and the second entity includes the third coefficient.

[0109] In some embodiments, the first calculation unit is also used to obtain a first relationship coefficient, a second coefficient, and a third coefficient; accumulate the first relationship coefficient, the second coefficient, and the third coefficient to obtain the sum of the relationship weights, wherein the relationship coefficients of each element entity with the target entity and all entities associated with the target entity include the sum of the relationship weights.

[0110] In some embodiments, the processing unit 54 further includes:

[0111] a second determining unit, configured to determine action label information corresponding to the user based on the acquired user behavior information, wherein the action label information includes preference information for characterizing historical operations of the user;

[0112] a third determining unit, coupled to the second determining unit, configured to determine, based on the preference information of the user's historical operations, category information corresponding to the user's historical operations, wherein the category information corresponding to the user's historical operations includes at least one of the following: operation device object information and operation time information;

[0113] The fourth determining unit is coupled to the third determining unit and is configured to determine a delivery strategy based on the classification information, wherein the delivery strategy includes a delivery device and / or delivery time for advertisement delivery.

[0114] In some embodiments, the acquisition module 52 is also used to obtain user portrait data corresponding to a preset user and text data corresponding to a preset scenario, wherein the user portrait data is generated based on user tag information and user behavior data, and the text data includes structured data and partially structured data; entities and entity attributes are identified and extracted from the user portrait data and text data respectively to obtain at least two entities and attribute information of each entity; knowledge fusion processing is performed on the relationship between any two entities and the attribute information of each entity to generate at least one entity pair relationship template, wherein the graph data includes entity pair relationship templates, and the knowledge fusion processing includes entity similarity calculation, deduplication, and disambiguation.

[0115] Figure 6 is a schematic structural diagram of an electronic device according to an embodiment of the present application. Figure 6As shown, an embodiment of the present application provides an electronic device, including a processor 61, a communication interface 62, a memory 63 and a communication bus 64, wherein the processor 61, the communication interface 62, and the memory 63 communicate with each other through the communication bus 64.

[0116] Memory 63, for storing computer programs;

[0117] Processor 61 is used to execute the program stored in memory 63 to achieve Figure 1 The method steps in .

[0118] Processing implementation in the server Figure 1 The technical effects of the method steps in the above embodiment are the same as those of the above embodiment. Figure 1 The technical effects of the advertising delivery method are the same and will not be repeated here.

[0119] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0120] The communication interface is used for communication between the above electronic device and other devices.

[0121] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor.

[0122] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0123] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the advertisement delivery method provided in any of the aforementioned method embodiments are implemented.

[0124] In another embodiment provided by the present application, a computer program product including instructions is also provided, which, when executed on a computer, enables the computer to execute the advertisement delivery method described in any one of the above embodiments.

[0125] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0126] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is intended to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. An advertisement delivery method, characterized in that: include: receiving an advertisement to be delivered, and determining advertisement element information corresponding to the advertisement, wherein the advertisement element information includes a plurality of element entities; Obtain graph data from a preset graph database, wherein the graph data is formed by performing entity extraction and knowledge fusion on the corresponding user portrait data; Determining a target user based on the relevance of the plurality of element entities to all entities associated with a target entity in the graph data, wherein the target entity is a named entity used to represent a user name; Determining a delivery strategy based on the user behavior information of the target user, and delivering the advertisement to the target user based on the delivery strategy; Specifically, the target user is determined based on the relevance of the plurality of element entities with all entities associated with the target entity in the graph data, including: extracting the target entity corresponding to the graph data and all entities associated with the target entity; sequentially calculating the relationship coefficients of each element entity with the target entity and all entities associated with the target entity, and determining the sum of the relationship coefficients of each element entity with the target entity and all entities associated with the target entity, wherein the relationship coefficient is used to characterize the degree of association between the corresponding entity and the element entity; judging whether the sum of the relationship coefficients is greater than a preset threshold; and determining that the user corresponding to the target entity is the target user when it is judged that the sum of the relationship coefficients is greater than the preset threshold; Specifically, the relationship coefficients between each of the element entities and the target entity and all entities associated with the target entity are calculated in sequence, including: determining a first entity among all entities associated with the target entity and a second entity associated with the first entity, wherein the first entity is an entity directly associated with the target entity; using natural language processing (NLP), calculating a first relationship coefficient between the element entity and the target entity, a second relationship coefficient between the target entity and the first entity, and a third relationship coefficient between the first entity and the second entity; based on the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient, determining the relationship coefficients between the element entity and the target entity and all entities associated with the target entity; Specifically, based on the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient, determining the relationship coefficient between the element entity and the target entity and all entities associated with the target entity includes: determining the first relationship coefficient as the relationship coefficient corresponding to the element entity and the target entity; multiplying the first relationship coefficient and the second relationship coefficient to obtain a second coefficient, wherein the relationship coefficient corresponding to the element entity and the first entity includes the second coefficient; multiplying the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient to obtain a third coefficient, and determining that the relationship coefficient corresponding to the element entity and the second entity includes the third coefficient; Specifically, determining the sum of the relationship coefficients between each of the element entities and the target entity and all entities associated with the target entity includes: obtaining the first relationship coefficient, the second coefficient and the third coefficient; accumulating the first relationship coefficient, the second coefficient and the third coefficient to obtain the sum of the relationship weights, wherein the sum of the relationship coefficients between each of the element entities and the target entity and all entities associated with the target entity includes the sum of the relationship weights.

2. The method according to claim 1, characterized in that Determine a delivery strategy based on the target user's user behavior information, including: Determining action tag information corresponding to the user based on the acquired user behavior information, wherein the action tag information includes preference information for characterizing the user's historical operations; Determining, based on the preference information of the user's historical operations, classification information corresponding to the user's historical operations, wherein the classification information corresponding to the user's historical operations includes at least one of the following: operation device object information and operation time information; The delivery strategy is determined based on the classification information, wherein the delivery strategy includes a delivery device and / or delivery time for advertisement delivery.

3. The method according to claim 1, characterized in that The process of constructing the atlas data includes: Obtain user portrait data corresponding to a preset user and text data corresponding to a preset scenario, wherein the user portrait data is generated based on user tag information and user behavior data, and the text data includes structured data and partially structured data; Identifying and extracting entities and entity attributes from the user portrait data and the text data respectively to obtain at least two entities and attribute information of each entity; Perform knowledge fusion processing on the relationship between any two entities and the attribute information of each entity to generate at least one entity-to-relationship template, wherein the graph data includes the entity-to-relationship template, and the knowledge fusion processing includes entity similarity calculation, deduplication, and disambiguation.

4. An advertising delivery device, characterized in that: include: A receiving module, configured to receive an advertisement to be delivered and determine advertisement element information corresponding to the advertisement, wherein the advertisement element information includes a plurality of element entities; An acquisition module is used to obtain graph data from a preset graph database, wherein the graph data is formed by performing entity extraction and knowledge fusion on the corresponding user portrait data; a determination module, configured to determine a target user based on the relevance of the plurality of element entities and all entities associated with a target entity in the graph data, wherein the target entity is a named entity used to represent a user name; a processing module, configured to determine a delivery strategy according to user behavior information of the target user, and deliver the advertisement to the target user based on the delivery strategy; The determination module is specifically used to: determine the target user according to the correlation between the multiple element entities and all entities associated with the target entity in the graph data, including: extracting the target entity corresponding to the graph data and all entities associated with the target entity; calculating the relationship coefficient of each element entity with the target entity and all entities associated with the target entity in turn, and determining the sum of the relationship coefficients of each element entity with the target entity and all entities associated with the target entity, wherein the relationship coefficient is used to characterize the degree of association between the corresponding entity and the element entity; judging whether the sum of the relationship coefficients is greater than a preset threshold; and determining that the user corresponding to the target entity is the target user when it is judged that the sum of the relationship coefficients is greater than the preset threshold; The determination module is further configured to sequentially calculate the relationship coefficients between each of the element entities and the target entity and all entities associated with the target entity, including determining a first entity among all entities associated with the target entity and a second entity associated with the first entity, wherein the first entity is an entity directly associated with the target entity; utilizing natural language processing (NLP) to calculate a first relationship coefficient between the element entity and the target entity, a second relationship coefficient between the target entity and the first entity, and a third relationship coefficient between the first entity and the second entity; and determining the relationship coefficients between the element entity and the target entity and all entities associated with the target entity based on the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient. The determining module is further configured to: determine, based on the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient, a relationship coefficient between the element entity and the target entity and all entities associated with the target entity, including: determining the first relationship coefficient as the relationship coefficient corresponding to the element entity and the target entity; multiplying the first relationship coefficient and the second relationship coefficient to obtain a second coefficient, wherein the relationship coefficient corresponding to the element entity and the first entity includes the second coefficient; multiplying the first relationship coefficient, the second relationship coefficient, and the third relationship coefficient to obtain a third coefficient, and determining that the relationship coefficient corresponding to the element entity and the second entity includes the third coefficient; The determination module is also used to: determine the sum of the relationship coefficients between each of the element entities and the target entity and all entities associated with the target entity, including: obtaining the first relationship coefficient, the second coefficient and the third coefficient; accumulating the first relationship coefficient, the second coefficient and the third coefficient to obtain the sum of the relationship weights, wherein the sum of the relationship coefficients between each of the element entities and the target entity and all entities associated with the target entity includes the sum of the relationship weights.

5. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory for storing computer programs; The processor is configured to implement the advertisement delivery method according to any one of claims 1 to 3 when executing the program stored in the memory.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the advertisement delivery method according to any one of claims 1 to 3 is implemented.

Images