A content distribution method, device, computer device and storage medium

Abstract

This disclosure relates to the field of computer technology, and more specifically, to a content distribution method, apparatus, computer device, and storage medium. The content distribution method includes: acquiring initial multimedia content matching each of multiple search cognitive dimensions; wherein the search cognitive dimension refers to a category dimension corresponding to the multimedia content capable of establishing search cognition; performing feature clustering on the initial multimedia content corresponding to each of the search cognitive dimensions to determine at least one feature condition corresponding to each of the search cognitive dimensions; and, based on the feature conditions, filtering target multimedia content from a multimedia content library that meets the feature conditions corresponding to each of the search cognitive dimensions, and distributing the target multimedia content.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more specifically, to a content distribution method, apparatus, computer device, and storage medium. Background Technology

[0002] Search engines are a common tool for users to obtain the information they need. However, users cannot predict in advance what kind of content a search engine can search for; they generally need to try using search engines multiple times to develop an understanding of them. This causes some users to miss many opportunities to use search engines to find the information they want, and also leads to a waste of search engine resources.

[0003] Therefore, how to help users better understand the information that search engines can search for and improve their search awareness is a question worth studying. Summary of the Invention

[0004] This disclosure provides at least one content distribution method, apparatus, and system.

[0005] In a first aspect, embodiments of this disclosure provide a content distribution method, including:

[0006] Based on multiple search cognitive dimensions, initial multimedia content matching each search cognitive dimension is obtained; the search cognitive dimension refers to the category dimension corresponding to the multimedia content that can establish search cognition.

[0007] For each of the search cognitive dimensions, the initial multimedia content is clustered to determine at least one feature condition corresponding to each of the search cognitive dimensions.

[0008] Based on the aforementioned feature conditions, target multimedia content that meets the feature conditions corresponding to each of the aforementioned search cognitive dimensions is selected from the multimedia content library, and the target multimedia content is then distributed.

[0009] In one implementation, the search cognition dimension includes: the search intention dimension;

[0010] The step of obtaining initial multimedia content matching each of the search cognitive dimensions includes:

[0011] Based on the post-view search conversion ratio corresponding to each candidate multimedia content in the multimedia content library, and the average post-view search conversion ratio corresponding to each multimedia content in the multimedia content library, from each candidate multimedia content, the candidate multimedia content with the post-view search conversion ratio greater than the average value is selected as the initial multimedia content matching the search intention dimension.

[0012] The "view-to-search conversion ratio" refers to the ratio between the number of times each user initiates a search for content related to the candidate multimedia content within a preset time period after watching the candidate multimedia content, and the number of times each user watches the candidate multimedia content.

[0013] In another implementation, the search cognition dimension includes: the demand type dimension;

[0014] The step of obtaining the initial multimedia content matching each of the search intent dimensions includes:

[0015] Based on the search type corresponding to each candidate multimedia content in the multimedia content library, the candidate multimedia content that matches the target search type is selected as the initial multimedia content; the search type refers to the request type corresponding to the search request initiated after reading the candidate multimedia content.

[0016] In another implementation, the search cognition dimension includes: an active behavior dimension;

[0017] The step of obtaining initial multimedia content matching each of the search cognitive dimensions includes:

[0018] Based on the domain category to which each candidate multimedia content and its corresponding search content belong in the multimedia content library, and the relevance between the candidate multimedia content and its corresponding search content, candidate multimedia content whose corresponding search content matches the domain category of the candidate multimedia content and whose relevance is lower than a set threshold is selected from the candidate multimedia content as the initial multimedia content; the search content refers to the multimedia content in the search results corresponding to the search request initiated after reading the candidate multimedia content.

[0019] In one implementation, the step of performing feature clustering on the initial multimedia content corresponding to each of the search cognition dimensions to determine at least one feature condition corresponding to the initial multimedia content with search cognition includes:

[0020] For each of the initial multimedia contents under the search cognitive dimension, the initial multimedia contents are classified to determine at least one multimedia content category under the search cognitive dimension;

[0021] Based on the consumption characteristics and post-view search conversion ratio of each initial multimedia content under each search cognitive dimension, determine the consumption characteristic threshold and post-view search conversion ratio threshold for each search cognitive dimension.

[0022] The multimedia content category, the consumption characteristic threshold, and the post-view search conversion ratio threshold are used as the characteristic conditions.

[0023] In one implementation, the distribution of the target multimedia content includes:

[0024] Based on the target multimedia content, target recommended keywords are determined;

[0025] The target recommended keywords corresponding to the target multimedia content are displayed on the target page; the target recommended keywords are displayed after being triggered.

[0026] In one implementation, the distribution of the target multimedia content includes:

[0027] The target multimedia content is displayed on the information feed recommendation page.

[0028] In one implementation, the distribution of the target multimedia content includes:

[0029] According to the priority order corresponding to each of the search cognitive dimensions, and / or the attribute characteristics of each of the target multimedia content, the target multimedia content corresponding to each of the search cognitive dimensions is sorted, and the target multimedia content is distributed based on the sorting results.

[0030] In one implementation, the step of filtering target multimedia content from the multimedia content library that meets the feature conditions corresponding to each of the search cognitive dimensions, and distributing the target multimedia content, includes:

[0031] Select multiple first target multimedia contents that meet the specified characteristics from the multimedia content library;

[0032] The plurality of first target multimedia content are input into the recommendation content model to obtain a plurality of second target multimedia content to be distributed; the recommendation content model is trained based on the distributed multimedia content samples and the consumption data after the distribution of the multimedia content samples.

[0033] The second target multimedia content is distributed.

[0034] In one implementation, determining target recommendation keywords based on the target multimedia content includes:

[0035] Based on the preset domain distribution requirements and / or genre distribution requirements, target multimedia content that meets the domain distribution requirements and / or genre distribution requirements is determined from each of the target multimedia content.

[0036] The target recommended words are determined based on the target multimedia content that meets the domain distribution requirements and / or genre distribution requirements.

[0037] Secondly, embodiments of this disclosure also provide a content distribution device, comprising:

[0038] The acquisition module is used to acquire initial multimedia content matching each of the multiple search cognitive dimensions; the search cognitive dimension refers to the category dimension that can establish search cognition.

[0039] The clustering module is used to perform feature clustering on the initial multimedia content corresponding to each search cognitive dimension, and determine at least one feature condition corresponding to each search cognitive dimension.

[0040] The content distribution module is used to filter target multimedia content that meets the feature conditions corresponding to each search cognitive dimension from the multimedia content library based on the feature conditions, and to distribute the target multimedia content.

[0041] Thirdly, embodiments of this disclosure also provide a computer device, including: a processor and a memory, the memory storing machine-readable instructions executable by the processor, the processor being configured to execute the machine-readable instructions stored in the memory, and when the machine-readable instructions are executed by the processor, the processor performs the steps of the content distribution method described in any of the preceding aspects.

[0042] Fourthly, embodiments of this disclosure also include a computer-readable storage medium storing a computer program, wherein when the computer program is run by a computer device, the computer device performs the steps of the content distribution method described in any of the preceding aspects.

[0043] The content distribution method, apparatus, computer device, and readable storage medium provided in this disclosure can acquire matching initial multimedia content for each of the multiple search cognition dimensions corresponding to multimedia content capable of establishing search cognition. The acquired initial multimedia content refers to multimedia content with search cognition under each search cognition dimension. Then, based on these initial multimedia content, the feature conditions corresponding to each search cognition dimension are summarized, i.e., feature clustering is performed on the initial multimedia content corresponding to each search cognition dimension. At least one feature condition corresponding to each search cognition dimension can be determined, and multimedia content that meets the feature conditions obtained from clustering can be considered to have search cognition. Subsequently, target multimedia content that meets the aforementioned feature conditions can be filtered from the multimedia content library. The target multimedia content obtained in this way is considered to bring search cognition, meaning that after reading this multimedia content, users can better understand the information direction that the search engine can search for. Therefore, when there is a relevant search need, they will consciously choose to use this search engine to find the content they want to know, thus helping users better use the search engine for information retrieval while improving the utilization rate of search engine resources.

[0044] To make the above-mentioned objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings used in the embodiments will be briefly described below. These drawings are incorporated in and constitute a part of this specification. They illustrate embodiments conforming to this disclosure and, together with the specification, serve to explain the technical solutions of this disclosure. It should be understood that the following drawings only show some embodiments of this disclosure and should not be considered as limiting the scope. Those skilled in the art can obtain other related drawings based on these drawings without creative effort.

[0046] Figure 1 A flowchart of a content distribution method provided by an embodiment of this disclosure is shown;

[0047] Figure 2 This is a method flow for determining matching feature conditions for each search cognitive dimension according to an embodiment of the present disclosure;

[0048] Figure 3 A schematic diagram illustrating the process of filtering and distributing target multimedia content according to an embodiment of this disclosure;

[0049] Figure 4 A schematic diagram illustrating the process of cognitive formation through search;

[0050] Figure 5 A schematic diagram of a content distribution device provided in an embodiment of this disclosure;

[0051] Figure 6 A schematic diagram of a computer device provided in an embodiment of this disclosure is shown. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. The components of the embodiments of this disclosure described and shown herein can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this disclosure is not intended to limit the scope of the claimed disclosure, but merely represents selected embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0053] Research has found that users cannot predict in advance what content a search engine can find; they generally need to try using search engines multiple times to develop an understanding of them. This causes some users to miss many opportunities to use search engines to find the information they want.

[0054] The shortcomings of the above solutions are the result of the inventor's practical experience and careful research. Therefore, the discovery process of the above problems and the solutions proposed in this disclosure below should be considered as the inventor's contribution to this disclosure.

[0055] Based on this, this disclosure provides a content distribution method that can acquire matching initial multimedia content for multiple search cognition dimensions corresponding to multimedia content that can establish search cognition; after performing feature clustering on the initial multimedia content corresponding to each search cognition dimension, at least one feature condition corresponding to each search cognition dimension can be determined; then, target multimedia content that meets the feature conditions can be filtered from the multimedia content library. The target multimedia content obtained in this way is considered to be multimedia content that can bring search cognition, that is, after reading this multimedia content, users can better understand the information direction that the search engine can search for, and thus, when they have relevant search needs, they will consciously choose to use this search engine to search for the content they want to know.

[0056] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0057] To facilitate understanding of this embodiment, a content distribution method disclosed in this disclosure will first be described in detail. The execution subject of the content distribution method provided in this disclosure is generally a computer device with certain computing power. The computer device may include, for example, a terminal device, a server, or other processing devices. In some possible implementations, the content distribution method can be implemented by the processor calling computer-readable instructions stored in the memory.

[0058] The content distribution method provided in the embodiments of this disclosure will be described below.

[0059] See Figure 1 The diagram shows a flowchart of a content distribution method provided in an embodiment of this disclosure. The method includes steps S101 to S103, wherein:

[0060] S101: Based on multiple search cognitive dimensions, obtain initial multimedia content that matches each of the search cognitive dimensions; the search cognitive dimension refers to the category dimension corresponding to the multimedia content that can establish search cognition.

[0061] Here, this embodiment of the disclosure mines multimedia content with search cognitive attributes from multiple dimensions, and refers to the dimensions that can bring about search cognitive attributes as search cognitive dimensions. Search cognitive attributes refer to a user's perception of the search capabilities of a search engine after reading multimedia content provided by the search engine. Specifically, multimedia content that can establish search cognitive attributes can be divided into multiple category dimensions according to different cognitive goals, thus obtaining multiple search cognitive dimensions.

[0062] For example, search cognition dimensions can be divided into search intention dimension, demand type dimension, and proactive behavior dimension. Their corresponding cognitive goals are: to stimulate short-term search motivation, to stimulate searches for expected content types (high-value types), and to stimulate long-term search intention. For details, please see the following introduction to these three search cognition dimensions.

[0063] In practice, the cognitive dimension of search can include the following three aspects:

[0064] (1) Search intention dimension;

[0065] The goal of this search cognitive dimension is to stimulate short-term search motivation, that is, after browsing relevant multimedia content, users are willing to search for more similar multimedia content in the short term.

[0066] Based on this, for this search intention dimension, we can start from the direction of "whether the post-view search conversion distribution has an advantage compared to the overall market", and filter the initial multimedia content under this search intention dimension. That is, we can select the content whose post-view search conversion ratio is higher than the average post-view search conversion ratio of each multimedia content in the multimedia content library.

[0067] Specifically, obtaining initial multimedia content matching each of the search cognitive dimensions includes:

[0068] Based on the post-view search conversion ratio of each candidate multimedia content in the multimedia content library, and the average post-view search conversion ratio of each multimedia content in the multimedia content library, the candidate multimedia content with a post-view search conversion ratio greater than the average value is selected from the candidate multimedia content as the initial multimedia content matching the search intention dimension.

[0069] The "view-to-search conversion rate" here refers to the ratio between the number of times each user initiates a search for content related to the candidate multimedia content within a preset time period after watching the candidate multimedia content, and the number of times each user watches the candidate multimedia content.

[0070] The candidate multimedia content here can be determined based on the content quality and / or consumption data of each multimedia content in the multimedia content library. For example, multimedia content in the multimedia content library with a content quality score greater than a set score and / or a consumption data volume greater than a set threshold can be selected as candidate multimedia content. The content quality score can be calculated based on a quality scoring model. During the training phase, the quality scoring model mainly considers the feature dimensions of the multimedia content, such as the content length and the amount of effective information. The consumption data volume can be calculated comprehensively based on the number of reads, the number of evaluations, etc.

[0071] In practical implementation, for each candidate multimedia content in the multimedia content library, the number of times each user initiates a search for related content within a preset time after viewing the candidate multimedia content can be counted. The ratio between the number of searches for related content and the number of times each user views the candidate multimedia content is used as the view-to-search conversion rate for that candidate multimedia content. Here, the number of times each user views the candidate multimedia content is also the number of visits to the candidate multimedia content; related content can be the same as the candidate multimedia content in the same field or multimedia content type (such as TV series or novels of the same theme, or news and information content under the same topic), or multimedia content that has a physical connection with the candidate multimedia content (such as TV series or movies starring the same star).

[0072] In addition, the average conversion rate of each candidate multimedia content reflects the overall average level of post-view search across the entire network. If the conversion rate of a candidate multimedia content is greater than the average, it can be considered that the post-view search conversion distribution of the candidate multimedia content has an advantage compared to the overall market, and thus the candidate multimedia content can be used as the initial multimedia content for matching search intent dimensions.

[0073] In addition, if there are a large number of candidate multimedia content with a conversion rate greater than the average after viewing, the above filtering conditions can be subject to stricter requirements. For example, candidate multimedia content with a conversion rate greater than the average after viewing or a ratio greater than a certain threshold can be filtered, or candidate multimedia content with a conversion rate greater than the average after viewing or a ratio exceeding a certain threshold can be filtered as the initial multimedia content matching the search intent dimension.

[0074] (2) Demand type dimension;

[0075] The characteristic of the demand type dimension is that the search type it brings is a high-value type; here, the search type refers to the type of request corresponding to the search request initiated after reading the candidate multimedia content, and the high-value type is the request type with high search cognitive value.

[0076] From the perspective of search results, search requests can be categorized into two types: 1) Result-ambiguous: Users do not know what the desired result should be before searching. This reflects the user's recognition of the search engine's comprehensive search capabilities. Examples include searching for information on trending topics or events, or searching for recipes or encyclopedic knowledge. 2) Result-explicit: Users know the result they need but simply need to find a specific one, such as directly entering the name of a live stream or novel.

[0077] For example, address-based requests are a type of request with explicit results. Users with address-based needs already know the result they need (e.g., account / video / resource / function / brand) and find it through search. For instance, when searching for a person / brand, they directly search for the name or its corresponding live stream / video; similarly, when searching for a specific resource (movie / anime / novel / game), they directly search for the resource name or a snippet; and when searching for a function (mini-game / plan / function entry), they directly search for the function, such as: creative inspiration, **mini-game entry**, etc. It's clear that users with address-based needs usually already know what they want, therefore, the cognitive value of this type of search request is relatively limited.

[0078] For example, news-related requests are a type of request with a certain time sensitivity and unclear results. Users with news-related needs learn about a hot topic or event from a search engine and then search for more information about hot topics or events. Therefore, multimedia content that triggers news-related search requests generally helps to form the perception of trending searches.

[0079] For example, general needs are a type of request with unclear results. Users with general needs tend to express their needs naturally when searching on search engines, aiming to find results that meet those needs. Examples include: Question-and-answer type: Where is the pet market in Pengjiang District? How to layout the HarmonyOS desktop?; Encyclopedia / Knowledge type: Toucan, Signs that a cold is almost gone; Guide / Tutorial type: Mainly focused on food recipes, talents and interests, games, and technology, such as how to make baked salt shrimp, beer chicken, lower body exercise for weight loss, singing while watching the stars, etc.; Recommendation / Review type: Mainly focused on e-commerce, life services, and general categories, such as recommendations for sweet romance novels, etc. General needs: The search information is relatively generalized, without a very clear need. It may be encyclopedia or a general category resource need, such as Chinese food, little tiger, shark, nursery rhymes, etc. Using general needs searches on search engines is an important indicator of a user's understanding of a search engine's comprehensive search capabilities. Multimedia content that can stimulate this type of search request has strong search cognitive value.

[0080] Based on the above analysis, in specific implementation, for the demand type dimension, the process of obtaining the initial multimedia content that matches the demand type dimension may include: based on the search type corresponding to each candidate multimedia content in the multimedia content library, selecting the candidate multimedia content that matches the target search type as the initial multimedia content.

[0081] Here, as can be seen from the above analysis, the goal of the demand type dimension is to stimulate search requests with high search cognitive value by reading multimedia content under this search cognitive dimension. Therefore, the search type after the target can be a demand type with high search cognitive value, such as the aforementioned information type with a certain timeliness and the long-term general demand type. Correspondingly, the initial multimedia content can include candidate multimedia content that can stimulate more of the aforementioned high-value search requests.

[0082] (3) Proactive Behavior Dimension;

[0083] The purpose of the proactive behavior dimension is to stimulate long-term search intentions; the search behavior following the initial multimedia content under the proactive behavior dimension is a generalized search behavior (that is, among the multimedia content searched after viewing, there is content in the same field but unrelated).

[0084] In specific implementation, under the active behavior dimension, obtaining initial multimedia content matching the active behavior dimension includes:

[0085] Based on the domain category to which each candidate multimedia content and its corresponding search content belong in the multimedia content library, and the relevance between the candidate multimedia content and its corresponding search content, candidate multimedia content whose corresponding search content matches the domain category of the candidate multimedia content and whose relevance is lower than a set threshold is selected from the candidate multimedia content and used as the initial multimedia content.

[0086] The aforementioned "search content after viewing" refers to the multimedia content in the search results corresponding to a search request initiated after reading the aforementioned candidate multimedia content.

[0087] Here, multimedia content that can lead to generalized search behavior is selected from among the candidate multimedia content. This means that after seeing the candidate multimedia content, the user has the search awareness to search for multimedia content in the same domain (i.e., they exhibit extended search behavior), rather than just searching for multimedia content that has a contextual or attribute-like relationship with the candidate multimedia content. For example, after viewing the candidate multimedia content "How to make tofu pudding sauce," a user searches for "recipe collection" and obtains multimedia content that belongs to the food domain but is not directly related to tofu pudding. This indicates that the user recognizes through the candidate multimedia content that the current search engine can search for multimedia content in the food domain. In this case, the candidate multimedia content can be used as the initial multimedia content.

[0088] S102: Perform feature clustering on the initial multimedia content corresponding to each search cognitive dimension to determine at least one feature condition corresponding to each search cognitive dimension.

[0089] In practice, after obtaining the initial multimedia content that matches each of the search cognition dimensions, the characteristics of the multimedia content corresponding to each search cognition dimension can be summarized based on these initial multimedia content, that is, at least one feature condition. Based on these feature conditions, more target multimedia content that can bring search cognition can be further filtered.

[0090] Generally, for each search cognition dimension, the characteristic conditions that each search cognition dimension meets can be summarized from multiple perspectives, such as multimedia content classification, consumption characteristics, and the conversion rate range after viewing and searching. For example... Figure 2 The diagram illustrates a method flow for determining matching feature conditions for each search cognitive dimension according to an embodiment of this disclosure, including:

[0091] S201: For the initial multimedia content under each search cognitive dimension, classify the initial multimedia content and determine at least one multimedia content category under the search cognitive dimension;

[0092] S202: Based on the consumption characteristics and post-view search conversion ratio of each initial multimedia content under each search cognitive dimension, determine the consumption characteristic threshold and post-view search conversion ratio threshold corresponding to each search cognitive dimension.

[0093] S203: The multimedia content category, the consumption feature threshold, and the post-view search conversion ratio threshold are used as the feature conditions.

[0094] In the above methodology, because the goals of content filtering differ under different search cognitive dimensions, the specific feature conditions corresponding to each search cognitive dimension are different. Specifically, at least one of the multimedia content category, consumption characteristic threshold, and post-view conversion ratio threshold differs between different search cognitive dimensions. For example, the goal of the search intention dimension is to stimulate short-term search motivation, aiming to identify target multimedia content that is intended to be searched more frequently in the short term. Therefore, the search intention dimension has a high requirement for the post-view conversion ratio. For example, the feature conditions corresponding to the search intention dimension include a high post-view conversion ratio: such as a post-view conversion ratio between 2% and 3% (the post-view conversion ratio threshold includes a minimum threshold and a maximum threshold; the maximum threshold is determined to filter out some multimedia content with abnormal traffic generation). In addition, it also includes multimedia content categories that are easy to stimulate short-term search motivation, such as medical or travel guide categories, as well as features related to consumption characteristics, such as the number of multimedia content visitors exceeding 100,000. Another example is the demand type dimension, which aims to stimulate high-value search requests. The corresponding multimedia content categories tend to be those with unclear results, such as food and technology. Furthermore, the threshold for the conversion rate after viewing to search corresponding to the demand type dimension is lower than the threshold for the conversion rate after viewing to search corresponding to the search intention dimension (the former's minimum threshold is lower than the latter's minimum threshold, and the former's maximum threshold is also lower than the latter's maximum threshold). The threshold for the number of users accessing multimedia content can be the same as or different from that of the search intention dimension. For example, the proactive behavior dimension focuses more on domain types that can lead to generalized search behavior, such as film and television or sports. Its corresponding threshold for the conversion rate after viewing to search can be between the thresholds for the conversion rates of the search intention and demand type dimensions, respectively. The threshold for the number of users accessing multimedia content can be the same as or different from those two dimensions.

[0095] S103: Based on the aforementioned feature conditions, select target multimedia content that meets the aforementioned feature conditions from the multimedia content library, and distribute the target multimedia content.

[0096] Here, based on the feature conditions determined for each search cognitive dimension, target multimedia content that meets the feature conditions is selected from the multimedia content library, and then the target multimedia content is distributed.

[0097] In one implementation, the target multimedia content can be displayed on the information feed push page. Specifically, on the information feed push page, the target multimedia content can be displayed together with multimedia content recommended based on other recommendation rules. The sorting rule could be to prioritize displaying the target multimedia content, or, if there are multiple target multimedia content items, to sort and display them according to their publication time, access popularity, etc.

[0098] In addition to the multimedia content itself, which carries search awareness, some search terms can also bring search awareness. Recommending search terms to users can remind them of the search capabilities of the search engine.

[0099] In practice, target recommendation keywords can be extracted from the target multimedia content. On the target page, the target recommendation keywords corresponding to the target multimedia content are displayed. When the target recommendation keyword is triggered, the corresponding target multimedia content is displayed.

[0100] Here, target recommendation words are extracted from the target multimedia content. Specifically, the target multimedia content can first be segmented to determine multiple keywords, and the relevance between the multiple keywords and the target multimedia content can be determined. The keyword with the highest relevance is selected as the target recommendation word.

[0101] The target page mentioned above can be any page in a search engine that can display recommended terms, such as the search page where a search request is initiated, the search results page, or the multimedia content details page in the search engine. Specifically, on the search page where a search is initiated, the target recommended terms corresponding to the target multimedia content are displayed. This can be done by displaying the target recommended terms in the search term recommendation area of ​​the search page, or by displaying the target recommended terms in the drop-down menu after triggering the search box. On the search results page, the target recommended terms can be displayed in a specific recommendation area outside the search results display area, or by inserting the target recommended terms into the search results. On the multimedia content details page, the target recommended terms can be displayed at the beginning or end of the multimedia content details page, or by inserting the target recommended terms into the multimedia content details page. By triggering the target recommended terms, users can see the target multimedia content associated with those recommended terms, thereby further enhancing their search awareness.

[0102] In specific implementation, target multimedia content that meets the preset domain distribution requirements and / or genre distribution requirements can be determined from each of the target multimedia content; and the target recommended words can be determined based on the target multimedia content that meets the domain distribution requirements and / or genre distribution requirements.

[0103] Here, due to the need to popularize search awareness of certain fields / genres (for example, users may not know that a certain search engine can search for recipes, or there may be a need to promote video and text content), or the need to distribute target recommended words as widely as possible across different fields / genres, this embodiment of the disclosure determines target multimedia content that meets the actual distribution needs from various target multimedia content based on the actual field distribution needs and / or genre distribution needs, and then obtains target recommended words based on these target multimedia content. In this way, the target recommended words obtained can also meet the distribution needs of fields or genres.

[0104] When distributing multiple target multimedia content, it is necessary to sort the multiple target multimedia content. In order to better leverage the search cognitive value of each target multimedia content, we can consider the different importance of different search cognitive dimensions to the search engine, as well as the differences in consumption data, quality scores, and other attribute characteristics of each target multimedia content.

[0105] In implementation, the target multimedia content corresponding to each search cognitive dimension can be sorted according to its priority order and / or the attribute characteristics of each target multimedia content, and then distributed based on the sorting results. For example, if the focus is on improving search engine motivation in the short term, the search intention dimension can be set as the highest priority; if the focus is on bringing more stable search traffic to the search engine, the demand type dimension can be set as the highest priority; and if the focus is on stimulating long-term search intention, the proactive behavior dimension can be set as the highest priority. In addition, target multimedia content with large amounts of consumer data and high quality scores can be prioritized. When combining the priority of search cognitive dimensions and the attribute characteristics of target multimedia content, the content can first be sorted according to the priority of the search cognitive dimensions, and then, for multiple target multimedia contents under each search cognitive dimension, sorted according to their attribute characteristics.

[0106] In practice, after selecting multimedia content that meets the aforementioned characteristics from the multimedia content library, if the number of multimedia content is large, further filtering can be performed to reduce the number of multimedia content pushed, allowing users to read the recommended multimedia content more targetedly and with higher quality, thereby improving the utilization efficiency of high-value multimedia content.

[0107] like Figure 3 The diagram illustrates a process for filtering and distributing target multimedia content according to an embodiment of this disclosure, including the following steps:

[0108] S301: Select multiple first target multimedia contents that meet the aforementioned characteristic conditions from the multimedia content library;

[0109] S302: Input the plurality of first target multimedia content into the recommendation content model to obtain a plurality of second target multimedia content to be distributed; the recommendation content model is trained based on the distributed multimedia content samples and the consumption data after the distribution of the multimedia content samples;

[0110] S303: Distribute the second target multimedia content.

[0111] In this embodiment, a recommendation content model is used to further filter the multimedia content to be distributed. This recommendation content model is trained based on multimedia content samples and the corresponding consumption data. The consumption data includes, for example, user access counts, discussion popularity, consumption duration, and number of reposts. Therefore, the recommendation content model trained based on this sample data can recommend target multimedia content to users that can generate high traffic and high consumption value.

[0112] The content distribution method provided in this embodiment can obtain matching initial multimedia content for each of the multiple search cognition dimensions corresponding to multimedia content capable of establishing search cognition. The initial multimedia content obtained here refers to the multimedia content with search cognition under each search cognition dimension. Then, based on these initial multimedia content, the characteristic conditions of the multimedia content with search cognition are summarized, that is, feature clustering is performed on the initial multimedia content corresponding to each search cognition dimension. This determines at least one characteristic condition corresponding to the initial multimedia content with search cognition under each search cognition dimension. Multimedia content that meets the characteristic conditions obtained from clustering can be considered to have search cognition. Subsequently, target multimedia content that meets the aforementioned characteristic conditions can be filtered from the multimedia content library. The target multimedia content obtained in this way is considered to bring search cognition, meaning that after reading this multimedia content, users can better understand the information direction that the search engine can search for. Therefore, when they have relevant search needs, they will consciously choose to use this search engine to find the content they want to know, thus helping users better use the search engine for information retrieval while improving the utilization rate of search engine resources.

[0113] like Figure 4The diagram illustrates the process of search cognition formation. A user sees target multimedia content or recommended keywords distributed on any page of a search engine. After becoming interested in a particular target multimedia content or recommended keyword, the user initiates a search request related to that target multimedia content or recommended keyword on the search engine. Upon obtaining matching search results, the search need is satisfied, thus forming cognition of the search engine's search capabilities, i.e., search cognition. As described in the above embodiments, the present disclosure provides target multimedia content and recommended keywords with high search cognition value, thereby helping to better form search cognition of the search engine and improving the utilization rate of search engine-related resources.

[0114] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0115] Based on the same inventive concept, this disclosure also provides a content distribution device corresponding to the content distribution method. Since the principle of the device in this disclosure for solving the problem is similar to that of the content distribution method described above, the implementation of the device can refer to the implementation of the content distribution method described above, and the repeated parts will not be described again.

[0116] Reference Figure 5 The diagram shown is a schematic representation of a content distribution device provided in an embodiment of this disclosure. The device includes: an acquisition module 51, a clustering module 52, and a content distribution module 53; wherein,

[0117] The acquisition module 51 is used to acquire initial multimedia content matching each of the multiple search cognition dimensions; the search cognition dimension refers to the category dimension corresponding to the multimedia content that can establish search cognition.

[0118] The clustering module 52 is used to perform feature clustering on the initial multimedia content acquired by the acquisition module 51 corresponding to each of the search cognition dimensions, and to determine at least one feature condition corresponding to each of the search cognition dimensions.

[0119] The content distribution module 53 is used to filter target multimedia content that meets the feature conditions corresponding to each search cognitive dimension from the multimedia content library based on the feature conditions determined by the clustering module 52 for each search cognitive dimension, and to distribute the target multimedia content.

[0120] In one implementation, the search cognition dimension includes: a search intention dimension; the acquisition module 51 is specifically used for:

[0121] Based on the post-view search conversion ratio corresponding to each candidate multimedia content in the multimedia content library, and the average post-view search conversion ratio corresponding to each multimedia content in the multimedia content library, candidate multimedia content with a post-view search conversion ratio greater than the average value is selected from the candidate multimedia content as the initial multimedia content matching the search intention dimension; wherein, the post-view search conversion ratio refers to the ratio between the number of times related content to the candidate multimedia content is searched within a preset time after watching the candidate multimedia content, and the number of times the candidate multimedia content is watched.

[0122] In another embodiment, the search cognition dimension includes: a demand type dimension; the acquisition module 51 is specifically used for:

[0123] Based on the search type corresponding to each candidate multimedia content in the multimedia content library, the candidate multimedia content that matches the target search type is selected as the initial multimedia content; the search type refers to the request type corresponding to the search request initiated after reading the candidate multimedia content.

[0124] In another embodiment, the search cognition dimension includes: an active behavior dimension; the acquisition module 51 is specifically used for:

[0125] Based on the domain category to which each candidate multimedia content and its corresponding search content belong in the multimedia content library, and the relevance between the candidate multimedia content and its corresponding search content, candidate multimedia content whose corresponding search content matches the domain category of the candidate multimedia content and whose relevance is lower than a set threshold is selected from the candidate multimedia content as the initial multimedia content; the search content refers to the multimedia content in the search results corresponding to the search request initiated after reading the candidate multimedia content.

[0126] In one implementation, clustering module 52 is specifically used for:

[0127] For each initial multimedia content under the search cognitive dimension, the initial multimedia content is categorized to determine at least one multimedia content category under the search cognitive dimension; based on the consumption characteristics and post-view search conversion ratio corresponding to each initial multimedia content under each search cognitive dimension, a consumption characteristic threshold and a post-view search conversion ratio threshold corresponding to each search cognitive dimension are determined; the multimedia content category, the consumption characteristic threshold, and the post-view search conversion ratio threshold are used as the feature conditions.

[0128] In one implementation, the content distribution module 53 is specifically used for:

[0129] Based on the target multimedia content, target recommendation keywords are determined; on the target page, the target recommendation keywords corresponding to the target multimedia content are displayed; after the target recommendation keywords are triggered, the corresponding target multimedia content is displayed.

[0130] In one implementation, the content distribution module 53 is specifically used to: display the target multimedia content on the information flow recommendation page.

[0131] In one implementation, the content distribution module 53 is specifically used for:

[0132] According to the priority order corresponding to each of the search cognitive dimensions, and / or the attribute characteristics of each of the target multimedia content, the target multimedia content corresponding to each of the search cognitive dimensions is sorted, and the target multimedia content is distributed based on the sorting results.

[0133] In one implementation, the content distribution module 53 is specifically used for:

[0134] Multiple first target multimedia contents that meet the aforementioned feature conditions are selected from the multimedia content library; the multiple first target multimedia contents are input into a recommendation content model to obtain multiple second target multimedia contents to be distributed; the recommendation content model is trained based on the distributed multimedia content samples and the corresponding post-distribution consumption data of the multimedia content samples; the second target multimedia contents are then distributed.

[0135] In one implementation, the content distribution module 53 is further configured to:

[0136] Based on preset domain distribution requirements and / or genre distribution requirements, target multimedia content that meets the domain distribution requirements and / or genre distribution requirements is determined from each of the target multimedia content; based on the target multimedia content that meets the domain distribution requirements and / or genre distribution requirements, the target recommended words are determined.

[0137] The processing flow of each module in the device and the interaction flow between each module can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0138] This disclosure also provides a computer device, such as... Figure 6 The diagram shown is a schematic representation of a computer device structure provided in an embodiment of this disclosure, including:

[0139] A processor 61 and a memory 62; the memory 62 stores machine-readable instructions executable by the processor 61, and the processor 61 executes the machine-readable instructions stored in the memory 62. When the machine-readable instructions are executed by the processor 61, the processor 61 performs the following steps:

[0140] Based on multiple search cognitive dimensions, initial multimedia content matching each search cognitive dimension is obtained; the search cognitive dimension refers to the category dimension corresponding to the multimedia content that can establish search cognition.

[0141] For each of the search cognitive dimensions, the initial multimedia content is clustered to determine at least one feature condition corresponding to each of the search cognitive dimensions.

[0142] Based on the aforementioned feature conditions, target multimedia content that meets the feature conditions corresponding to each of the aforementioned search cognitive dimensions is selected from the multimedia content library, and the target multimedia content is then distributed.

[0143] The aforementioned memory 62 includes a main memory 621 and an external memory 622. The main memory 621, also known as internal memory, is used to temporarily store the computational data in the processor 61, as well as the data exchanged with external memory such as a hard disk. The processor 61 exchanges data with the external memory 622 through the main memory 621.

[0144] The specific execution process of the above instructions can be referred to the steps of the content distribution method described in the embodiments of this disclosure, and will not be repeated here.

[0145] This disclosure also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the content distribution method described in the above method embodiments. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0146] This disclosure also provides a computer program product carrying program code. The program code includes instructions that can be used to execute the steps of the content distribution method described in the above method embodiments. For details, please refer to the above method embodiments, which will not be repeated here.

[0147] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0148] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this disclosure, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0149] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0150] In addition, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0151] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0152] Finally, it should be noted that the above-described embodiments are merely specific implementations of this disclosure, used to illustrate the technical solutions of this disclosure, and not to limit it. The protection scope of this disclosure is not limited thereto. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this disclosure. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be covered within the protection scope of this disclosure. Therefore, the protection scope of this disclosure should be determined by the protection scope of the claims.

Claims

1. A content distribution method, characterized in that, include: Based on multiple search cognitive dimensions, obtain initial multimedia content that matches each of the search cognitive dimensions; The search cognition dimension refers to the category dimension corresponding to the multimedia content that enables search cognition. The search cognition is the user's cognition of the search capabilities of the search engine after reading the multimedia content provided by the search engine. For each of the search cognitive dimensions, the initial multimedia content is clustered to determine at least one feature condition corresponding to each of the search cognitive dimensions. Based on the aforementioned feature conditions, target multimedia content that meets the feature conditions corresponding to each of the aforementioned search cognitive dimensions is selected from the multimedia content library, and the target multimedia content is then distributed.

2. The method according to claim 1, characterized in that, The search cognition dimension includes: the search intention dimension; The step of obtaining initial multimedia content matching each of the search cognitive dimensions includes: Based on the post-view search conversion ratio corresponding to each candidate multimedia content in the multimedia content library, and the average post-view search conversion ratio corresponding to each multimedia content in the multimedia content library, from each candidate multimedia content, the candidate multimedia content with the post-view search conversion ratio greater than the average value is selected as the initial multimedia content matching the search intention dimension. The "view-to-search conversion ratio" refers to the ratio between the number of times the candidate multimedia content is searched for within a preset time period after viewing the candidate multimedia content, and the number of times the candidate multimedia content is viewed.

3. The method according to claim 1, characterized in that, The search cognitive dimensions include: the dimension of demand type; The step of obtaining initial multimedia content matching each of the search cognitive dimensions includes: Based on the search type corresponding to each candidate multimedia content in the multimedia content library, the candidate multimedia content that matches the target search type is selected as the initial multimedia content; the search type refers to the request type corresponding to the search request initiated after reading the candidate multimedia content.

4. The method according to claim 1, characterized in that, The search cognition dimension includes: the proactive behavior dimension; The step of obtaining initial multimedia content matching each of the search cognitive dimensions includes: Based on the domain category to which each candidate multimedia content and its corresponding search content belong in the multimedia content library, and the relevance between the candidate multimedia content and its corresponding search content, candidate multimedia content whose corresponding search content matches the domain category of the candidate multimedia content and whose relevance is lower than a set threshold is selected from the candidate multimedia content as the initial multimedia content; the search content refers to the multimedia content in the search results corresponding to the search request initiated after reading the candidate multimedia content.

5. The method according to claim 1, characterized in that, The step of performing feature clustering on the initial multimedia content corresponding to each of the search cognition dimensions to determine at least one feature condition corresponding to the initial multimedia content with search cognition includes: For each of the initial multimedia contents under the search cognitive dimension, the initial multimedia contents are classified to determine at least one multimedia content category under the search cognitive dimension; Based on the consumption characteristics and post-view search conversion ratio of each initial multimedia content under each search cognition dimension, determine the consumption characteristic threshold and post-view search conversion ratio threshold corresponding to each search cognition dimension. The multimedia content category, the consumption characteristic threshold, and the post-view search conversion ratio threshold are used as the characteristic conditions.

6. The method according to claim 1, characterized in that, The distribution of the target multimedia content includes: Based on the target multimedia content, target recommended keywords are determined; The target recommended keywords corresponding to the target multimedia content are displayed on the target page; the target recommended keywords are displayed after being triggered.

7. The method according to claim 1, characterized in that, The distribution of the target multimedia content includes: The target multimedia content is displayed on the information feed recommendation page.

8. The method according to claim 1, characterized in that, The distribution of the target multimedia content includes: According to the priority order corresponding to each of the search cognitive dimensions, and / or the attribute characteristics of each of the target multimedia content, the target multimedia content corresponding to each of the search cognitive dimensions is sorted, and the target multimedia content is distributed based on the sorting results.

9. The method according to claim 1, characterized in that, The step of selecting target multimedia content from the multimedia content library that meets the feature conditions corresponding to each of the search cognitive dimensions, and distributing the target multimedia content, includes: Select multiple first target multimedia contents that meet the specified characteristics from the multimedia content library; The plurality of first target multimedia content are input into the recommendation content model to obtain a plurality of second target multimedia content to be distributed; the recommendation content model is trained based on the distributed multimedia content samples and the consumption data after the distribution of the multimedia content samples. The second target multimedia content is distributed.

10. The method according to claim 6, characterized in that, Based on the target multimedia content, target recommended keywords are determined, including: Based on the preset domain distribution requirements and / or genre distribution requirements, target multimedia content that meets the domain distribution requirements and / or genre distribution requirements is determined from each of the target multimedia content. The target recommended words are determined based on the target multimedia content that meets the domain distribution requirements and / or genre distribution requirements.

11. A content distribution device, characterized in that, include: The acquisition module is used to acquire initial multimedia content that matches each of the multiple search cognitive dimensions. The search cognition dimension refers to the category dimension that enables the establishment of search cognition. The search cognition is the user's cognition of the search capabilities of the search engine after reading multimedia content provided by the search engine. The clustering module is used to perform feature clustering on the initial multimedia content corresponding to each search cognitive dimension, and determine at least one feature condition corresponding to each search cognitive dimension. The content distribution module is used to filter target multimedia content that meets the feature conditions corresponding to each search cognitive dimension from the multimedia content library based on the feature conditions, and to distribute the target multimedia content.

12. A computer device, characterized in that, include: A processor and a memory, the memory storing machine-readable instructions executable by the processor, the processor executing the machine-readable instructions stored in the memory, wherein when the machine-readable instructions are executed by the processor, the processor performs the steps of the content distribution method as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a computer device, performs the steps of the content distribution method as described in any one of claims 1 to 10.

Images