Tagging information generation method and device, computer device, and storage medium

Abstract

The application relates to a label information generation method and device, computer equipment and a storage medium. The method comprises the following steps: when the watching record of a target object on a current video platform does not reach a preset condition, obtaining the historical watching record of the target object on other video platforms, the historical watching record comprising video information of each video watched and corresponding watching behavior information; extracting video labels of each video from the video information of each video; determining the weight of each video label according to the video information and the corresponding watching behavior information of each video under each video label; and generating label information of the target object according to each video label and the weight, and recommending videos for the target object based on the label information by the current video platform. The method can more accurately predict the interest tendency of the target object.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer device, and storage medium for generating tagged information. Background Technology

[0002] User profiling, or user information tagging, involves businesses collecting and analyzing data on key consumer information such as social attributes, lifestyle habits, and consumption behaviors to abstract a complete commercial profile of each user. This is a fundamental way for businesses to apply big data technology. In the era of big data, user information permeates the internet. Abstracting each specific piece of user information into tags and using these tags to concretize the user's image—that is, generating a user profile—allows for targeted services to be provided. Tags are typically highly refined characteristic identifiers, such as an age group tag: 25-35 years old, or a geographical tag: Beijing.

[0003] Short video platforms typically recommend videos to users based on user profiles. These user profiles, as the primary input to the recommendation system, are crucial to the recommendation effectiveness. To establish personalized features, recommendation systems assign tens of thousands or even tens of millions of tags to users, including preferences and preference intensity values. For example, a user's profile might look like this: Figure 1 As shown, probability is used to represent credibility; a higher probability can be understood as higher credibility. Preference value is used to represent liking; a higher preference value can be understood as liking more.

[0004] In traditional technologies, short video platforms generate user profiles based on in-app historical behavior records. However, for new users, the lack of in-app historical behavior records means that predictions of user interests can only be made based on basic user attributes (such as age, gender, and city), which has low accuracy and cannot effectively solve the user cold start problem. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, computer device, and storage medium for generating tagged information that can more accurately predict users' interests and preferences, in order to address the aforementioned technical problems.

[0006] A method for generating tagged information, the method comprising:

[0007] When the target object's viewing history on the current video platform does not meet the preset conditions, the target object's historical viewing history on other video platforms is obtained. The historical viewing history includes video information of each video watched and corresponding viewing behavior information.

[0008] Extract video tags from the video information of each video;

[0009] The weight of each video tag is determined based on the video information and corresponding viewing behavior information of each video under each video tag;

[0010] Based on the video tags and their weights, tagged information of the target object is generated, and the current video platform recommends videos for the target object based on the tagged information.

[0011] A tagging information generation device, the device comprising:

[0012] The acquisition module is used to acquire the target object's historical viewing records on other video platforms when the target object's viewing records on the current video platform do not meet the preset conditions. The historical viewing records include video information of each video watched and corresponding viewing behavior information.

[0013] The extraction module is used to extract video tags from the video information of each video.

[0014] The determination module is used to determine the weight of each video tag based on the video information of each video under each video tag and the corresponding viewing behavior information;

[0015] The generation module is used to generate tagged information of the target object based on each video tag and its weight, and the current video platform recommends videos for the target object based on the tagged information.

[0016] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program performing the following steps:

[0017] When the target object's viewing history on the current video platform does not meet the preset conditions, the target object's historical viewing history on other video platforms is obtained. The historical viewing history includes video information of each video watched and corresponding viewing behavior information.

[0018] Extract video tags from the video information of each video;

[0019] The weight of each video tag is determined based on the video information and corresponding viewing behavior information of each video under each video tag;

[0020] Based on the video tags and their weights, tagged information of the target object is generated, and the current video platform recommends videos for the target object based on the tagged information.

[0021] A computer-readable storage medium having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0022] When the target object's viewing history on the current video platform does not meet the preset conditions, the target object's historical viewing history on other video platforms is obtained. The historical viewing history includes video information of each video watched and corresponding viewing behavior information.

[0023] Extract video tags from the video information of each video;

[0024] The weight of each video tag is determined based on the video information and corresponding viewing behavior information of each video under each video tag;

[0025] Based on the video tags and their weights, tagged information of the target object is generated, and the current video platform recommends videos for the target object based on the tagged information.

[0026] A computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0027] When the target object's viewing history on the current video platform does not meet the preset conditions, the target object's historical viewing history on other video platforms is obtained. The historical viewing history includes video information of each video watched and corresponding viewing behavior information.

[0028] Extract video tags from the video information of each video;

[0029] The weight of each video tag is determined based on the video information and corresponding viewing behavior information of each video under each video tag;

[0030] Based on the video tags and their weights, tagged information of the target object is generated, and the current video platform recommends videos for the target object based on the tagged information.

[0031] The aforementioned tagging information generation method, apparatus, computer equipment, computer-readable storage medium, and computer program product extract video tags and obtain the weights of each video tag from the target object's historical viewing records on other video platforms to generate tagged information for the target object. This can improve profile coverage, solve the current user cold start problem on video platforms, and the video tags are more suitable as user interest tags, which helps to obtain a more accurate user profile and thus make more precise video recommendations. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of a user profile in one embodiment;

[0033] Figure 2This is a diagram illustrating the application environment of a tagging information generation method in one embodiment.

[0034] Figure 3 This is a flowchart illustrating a method for generating tagged information in one embodiment;

[0035] Figure 4 This is a flowchart illustrating the steps for obtaining the historical viewing records of a target object on other video platforms in one embodiment.

[0036] Figure 5 This is a schematic diagram of the process for obtaining historical viewing records in one embodiment;

[0037] Figure 6 This is a schematic diagram of the user profile results in one embodiment;

[0038] Figure 7 This is a flowchart illustrating the steps of determining the weight of each video tag based on the video information and corresponding viewing behavior information of each video under each video tag in one embodiment.

[0039] Figure 8 This is a flowchart illustrating the steps of obtaining the viewing factor of a video based on the viewing duration and viewing time in one embodiment.

[0040] Figure 9 This is a schematic diagram of the normalized function curve in one embodiment;

[0041] Figure 10 This is a schematic diagram of a friend relationship chain in one embodiment;

[0042] Figure 11 This is a flowchart illustrating a portrait generation method in one embodiment;

[0043] Figure 12 This is a flowchart illustrating a portrait generation method in one embodiment;

[0044] Figure 13 This is a structural block diagram of an image generation device in one embodiment;

[0045] Figure 14 This is an internal structural diagram of a computer device in one embodiment;

[0046] Figure 15 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0048] The tagging information generation method provided in this application can be applied to, for example... Figure 2 In the application environment shown, terminal 202 communicates with first server 204 and second server 206 via a network, and first server 204 and second server 206 communicate via the network. Users can access short video platforms through terminal 202, with first server 204 being the server hosting the short video platform. Users can also access other video platforms through terminal 202, with second server 206 being the server hosting the other video platform. Accounts of the same user on the short video platform and other video platforms can be linked, for example, by binding the same user identifier. First server 204 can obtain behavioral data of linked accounts on other video platforms from second server 206 to generate tagged information (i.e., user profiles) for the recommendation system applied to the short video platform. Terminal 202 includes, but is not limited to, mobile phones, computers, smart voice interaction devices, smart home appliances, and in-vehicle terminals. Both first server 204 and second server 206 can be implemented using independent servers or a server cluster consisting of multiple servers.

[0049] In one embodiment, such as Figure 3 As shown, a method for generating tagged information is provided, which can be applied to... Figure 2 The first server 204 in the example is used for illustration, including the following steps S302 to S308.

[0050] S302, when the target object's viewing history on the current video platform does not meet the preset conditions, obtain the target object's historical viewing history on other video platforms. The historical viewing history includes video information of each video watched and corresponding viewing behavior information.

[0051] The current video platform refers to a short video platform, and the target audience refers to users of that short video platform. Preset conditions may include, but are not limited to, viewing time exceeding a certain set duration or the number of videos viewed exceeding a certain set number. If the target audience's viewing history on the current video platform does not meet the preset conditions, it can be considered that the target audience is a new user of that platform, with little or no behavioral record, insufficient to reflect their interests. Therefore, the target audience's historical viewing history on other video platforms is used to predict their interests.

[0052] Other video platforms can be any other video platform that provides video viewing services. All user actions on these platforms (such as launching the application, starting to play a video, ending playback, pausing playback, etc.) can be reported through event tracking to generate historical viewing records. Obtaining the target audience's historical viewing records on other video platforms can specifically mean obtaining the target audience's historical viewing records on other video platforms over a recent period (such as the last month, three months, or six months). Based on this, the target audience's interests can be predicted more accurately.

[0053] S304 extracts the video tags of each video from the video information of each video.

[0054] Video tags represent extracted video features used to reflect the interests and preferences of the target audience watching the video. It's important to note that different videos can share the same video tag, meaning multiple videos can be included under the same tag. For example, for a 20-episode TV series, the video tag for each episode could be the title of the series.

[0055] S306. Determine the weight of each video tag based on the video information and corresponding viewing behavior information of each video under each video tag.

[0056] The weight of a video tag is used to characterize the target audience's interest, preference, or demand for videos under that tag. Multiple video tags can be obtained by analyzing video information from videos watched by the target audience on other video platforms, each tag corresponding to a weight. In one embodiment, a higher weight indicates that the target audience is more interested in, likes, or urgently wants to watch videos under that tag.

[0057] The weight of a video tag is related to the video information and the corresponding viewing behavior information. For example, if a video 'a' under video tag A and a video 'b' under video tag B have the same duration, and the target audience watches video 'a' for a longer duration than video 'b', it can be assumed that the target audience's interests are more inclined towards video 'a', and thus the weight of video tag A may increase accordingly. Similarly, if the target audience watches video 'a' more often than video 'b', it can also be assumed that the target audience's interests are more inclined towards video 'a', and thus the weight of video tag A may increase accordingly.

[0058] S308 generates tagged information for the target object based on each video tag and its weight, and the current video platform recommends videos for the target object based on the tagged information.

[0059] The tagged information of the target object can be understood as the target object's profile. The target object's profile is specified through various video tags. After obtaining the target object's profile, the current video platform can recommend more targeted videos that are more in line with the target object's interests based on this profile.

[0060] In the above-mentioned method for generating tagged information, video tags are extracted and their weights are obtained from the target object's historical viewing records on other video platforms. These tags are then used to generate tagged information for the target object, which can improve profile coverage and solve the current user cold start problem on video platforms. At the same time, video tags are more suitable as user interest tags, which helps to obtain more accurate user profiles and thus more precise video recommendations.

[0061] In one embodiment, such as Figure 4 As shown, the steps for obtaining the target object's historical viewing records on other video platforms may specifically include the following steps S402 to S406.

[0062] S402, Identify the accounts associated with the target on other video platforms and obtain the historical viewing records of those accounts on other video platforms.

[0063] Accounts of a target on the current video platform and other video platforms can be linked by binding the same user identifier. This allows the identification of the target's accounts on other video platforms using their current account and the linked user identifier. Once the linked accounts on other video platforms are identified, their associated behavioral data can be queried to obtain their historical viewing records on those platforms.

[0064] S404, When it is determined that an account corresponds to a user based on the login devices of the account on other video platforms, the historical viewing records of the account on other video platforms shall be used as the historical viewing records of the target on other video platforms.

[0065] An account can correspond to one user or multiple users. For example, if a video platform supports one account logging in on multiple devices simultaneously, then one account can correspond to multiple users (e.g., family members sharing a VIP account). Specifically, the login device information for an account on other video platforms can include the device type and the number of devices. Each device has a unique device identifier, and the device type and number are obtained by detecting the device identifier. In this embodiment, device type refers to the type of mobile device, including but not limited to mobile phones and tablets, and device number refers to the number of mobile devices.

[0066] In one embodiment, when the number of mobile devices used by an account to log in to other video platforms within a recent preset time period exceeds a first threshold, the account is determined to belong to multiple users. Both the recent preset time period and the first threshold can be set according to actual circumstances. For example, if the recent preset time period is set to the last three months and the first threshold is set to 5, then if an account logs in to other video platforms using more than 5 different mobile devices within the last three months, the account is considered to belong to multiple users.

[0067] In one embodiment, when the number of times an account logs into other video platforms on the same type of mobile device within a recent preset time period exceeds a second threshold, the account is determined to belong to multiple users. The recent preset time period and the second threshold can be set according to actual circumstances. For example, if the recent preset time period is set to the last three months and the second threshold is set to 2, then if an account logs into other video platforms using more than two different mobile phones or different tablets within the last three months, the account is considered to belong to multiple users.

[0068] When the login devices of an account on other video platforms do not meet the conditions for determining that an account corresponds to multiple users as described in the above embodiments, the account is determined to correspond to one user. In this case, the historical viewing records of the account on other video platforms are used as the historical viewing records of the target object on other video platforms. The same user may watch videos on both mobile devices (e.g., mobile phones) and non-mobile devices (e.g., PCs). Therefore, using the account-level profile as the final device profile can improve the profile coverage.

[0069] S406 When it is determined that an account corresponds to multiple users based on the login devices of the account on other video platforms, obtain the device identifier of the target object to log in to the current video platform, and filter out the historical viewing records associated with the device identifier from the historical viewing records of the account on other video platforms as the historical viewing records of the target object on other video platforms.

[0070] When the login devices of an account on other video platforms meet any of the conditions used in the above embodiments to determine that an account corresponds to multiple users, it is determined that the account corresponds to multiple users. At this time, from the historical viewing records of the account on other video platforms, the historical viewing records associated with the device identifier of the target object logging into the current video platform are filtered out as the historical viewing records of the target object on other video platforms. Thus, using the device-level profile as the final device profile can improve the accuracy of the profile.

[0071] In one embodiment, such as Figure 5 As shown, a process for obtaining historical viewing records is provided, including the following steps S502 to S510.

[0072] S502, Obtain Event Tracking Reported Data. Event tracking reported data can include all behavioral data of the target object on other video platforms, such as launching the application, starting to play the video, ending to play the video, pausing playback, and other behavioral data.

[0073] S504, perform data cleaning on the obtained data reported by the tracking points. In this embodiment, only the relevant data corresponding to the end of video playback can be retained.

[0074] S506, extracting viewing data from the past month.

[0075] S508 obtains viewing data at the device level based on viewing data and device identifier, which serves as the historical viewing record when an account corresponds to multiple users.

[0076] S510 obtains viewing data at the account level based on viewing data and account identifiers, which serves as the historical viewing record when an account corresponds to a user.

[0077] In the above embodiments, the mapping relationship between account, device, and user is identified. For different mapping situations, user profiles are mined from the account dimension and the device dimension, which helps to improve the coverage and accuracy of the profiles.

[0078] In one embodiment, the video information includes the video category and the video title; the step of extracting the video tag of each video from the video information of each video may specifically include: when the video category of each video is TV series, movie, variety show or animation, the video title of each video is used as the video tag of each video.

[0079] TV dramas, movies, variety shows, and anime videos are popular video types. Using the titles of these videos as video tags can help uncover more precise user profiles. For example... Figure 6 The diagram illustrates the user profile results in one embodiment. It should be noted that, in addition to profiles related to TV dramas, movies, variety shows, and animation, profiles for vertical categories such as sports and music can also be mined using similar methods.

[0080] In one embodiment, video information includes video duration, and viewing behavior information includes the number of views and the viewing duration and time of each view.

[0081] Users may watch the same video multiple times; the more times a user watches, the higher their level of interest in the video is considered to be. Each watch corresponds to a viewing duration and a viewing time. The video's total length can be understood as its inherent duration, and the viewing duration can be understood as the amount of time the video has been played when the user ends playback. The viewing duration is less than or equal to the video's total length. The viewing time can be the start time (the time when the user begins watching the video), the end time (the time when the user ends watching the video), or any time between the start and end times. This viewing time can be measured in days.

[0082] In one embodiment, such as Figure 7 As shown, the steps for determining the weight of each video tag based on the video information and corresponding viewing behavior information of each video under each video tag may specifically include the following steps S702 to S704.

[0083] S702: For each video under each video tag, for each viewing, obtain the viewing factor of the video under this viewing based on the viewing duration and viewing time of the video under this viewing.

[0084] Viewing factors are used to measure a user's interest in a video. In one embodiment, such as... Figure 8 As shown, the steps to obtain the viewing factor of a video under this viewing based on the viewing duration and viewing time can specifically include the following steps S802 to S808.

[0085] S802, based on the video's viewing duration in this viewing, obtain the video's viewing duration factor in this viewing.

[0086] Specifically, the viewing duration of the video in this viewing session is normalized to obtain a viewing duration factor for the video in this viewing session. The purpose of normalization is to map the variable to the range of 0 to 1. The normalization function can be the tanh function, and the tanh function curve is shown below. Figure 9 As shown, the advantage is that when the x-axis is very small, the curve slope is large, providing good discrimination; when the x-axis exceeds 1.5, the curve begins to flatten. In implementation, the variable to be normalized can be appropriately amplified or reduced before using the tanh function for normalization, for example, tanh(x*α), where x represents the variable and α represents the amplification or reduction coefficient.

[0087] S804: Based on the ratio of the viewing time to the video duration in this viewing session, obtain the video completion rate factor in this viewing session.

[0088] Specifically, the completion rate of the video in this viewing session is obtained by comparing the viewing duration to the total video length. This completion rate is then normalized to obtain a completion rate factor. The purpose of normalization is to map the variable to the range of 0 to 1; the tanh function can be used for normalization.

[0089] S806: Based on the time difference between the viewing time of the video in this viewing and the current time, obtain the time decay factor of the video in this viewing.

[0090] Specifically, the time decay factor of the video during this viewing is β. t-t0 Where β is the decay coefficient, 0 < β < 1; t represents the current time, i.e., the time required to generate the profile, in days; t0 represents the viewing time of the video in this viewing session, in days. As time progresses, it can be assumed that the user's interest in the video gradually decreases.

[0091] S808: Based on the video's viewing duration factor, completion rate factor, and time decay factor during this viewing, obtain the video's viewing factor for this viewing.

[0092] Specifically, the formula for calculating the viewing factor (denoted by V) of a video in this viewing context can be as follows:

[0093]

[0094] Where W represents the viewing duration factor, m represents the viewing duration factor weight, F represents the completion rate factor, n represents the completion rate factor weight, and β t-t0 This represents the time decay factor.

[0095] S704 integrates the viewing factors of each video under the video tag at each viewing time to obtain the weight of the video tag.

[0096] Specifically, the viewing factors of each video under a video tag are merged across different viewing sessions to obtain a merged viewing factor. This merged viewing factor is then normalized to obtain the weights of the video tags. The tanh function can be used for normalization. The formula for calculating the merged viewing factor (denoted by A) is as follows:

[0097]

[0098] Where n represents the total number of views of all videos under the video tag, and V i Let represent the viewing factor for the i-th time.

[0099] In the above embodiments, combining factors such as video duration, completion rate, number of views, and time decay to calculate the weight of video tags helps to obtain a more accurate user profile.

[0100] In one embodiment, when the number of video tags does not meet the quantity requirement, the associated object information of the target object is obtained, including the relationship closeness between the target object and each associated object; based on the relationship closeness between the target object and each associated object, as well as the tags and weights of each associated object, the propagation tag and weight of the target object are obtained.

[0101] The number of video tags does not meet the requirement, specifically, the number of video tags is less than the preset number. The preset number can be set according to actual needs and is not limited here. Associated objects can include objects that have a bidirectional relationship with the target object, specifically, the target object's friends.

[0102] The target audience may lack behavioral data on other video platforms, resulting in a lack of or very few video tags. When the number of video tags is less than a preset limit, the target audience's friend network can be used to infer their potential interests based on the tags their friends possess. For example, if most of the target audience's friends possess a certain tag, then it can be assumed that the target audience is more likely to possess that tag; if the target audience has a very close relationship with their friends, then it can be assumed that the target audience is more likely to possess that friend's tag.

[0103] Specifically, a friend relationship graph can be created, such as... Figure 10 As shown, each user is represented as a vertex, and friendship relationships are represented as edges. The edge weights represent the intimacy of the relationship between users. For example, w(A,B) represents the intimacy of the relationship between user A and user B, and its calculation formula is as follows:

[0104]

[0105] Where A∩B represents the number of mutual friends between user A and user B.

[0106] like Figure 10 As shown, each user's tag and tag weight are stored as attributes within that vertex. For example, user B has the tag `tag1` and the corresponding tag weight `weight1`. In each tag propagation, each vertex receives the attributes of its neighboring nodes, aggregating the tags and tag weights of its friends as its own propagated tag. Specifically, the propagated tag and its weight of node a can be calculated as follows:

[0107]

[0108] Where, weight a(tag) represents the weight of the propagated tag of node a; node j represents the j-th friend of node a; n represents the number of friends; w aj This represents the closeness of the relationship between node a and node j. For example... Figure 10 As shown, user A can obtain two propagation tags, tag1 and tag2. The weight of tag1 is w(A,B)*weight1, and the weight of tag2 is w(A,C)*weight2. If user C also has tag1 with a weight of weight4, then user A's tag1 weight is w(A,B)*weight1 + w(A,C)*weight4. After obtaining the weights using the above formula, the tanh function is used to normalize the calculated weights.

[0109] The above embodiments provide a graph-based tag propagation algorithm that is simple and intuitive. Furthermore, algorithms such as community detection, random walk, PageRank, and node embedding can be used to mine more tags; the specific algorithm can be selected based on the specific business requirements.

[0110] The step of generating tagged information for a target object based on video tags and their weights can specifically include: generating tagged information for the target object based on video tags and their weights, as well as the target object's propagation tags and their weights. Based on this, a profile of the target object is jointly determined by combining the video tags and their weights, and the propagation tags and their weights.

[0111] In the above embodiments, introducing a friend relationship chain for tag propagation can recall more tags, which helps to identify the potential interests of the target audience and further improve the profile coverage. It should be noted that, in addition to introducing a friend relationship chain for tag propagation to achieve tag generalization, tag generalization can also be based on tag content, such as calculating the similarity of TV series or using item-based collaborative filtering algorithms.

[0112] In one embodiment, such as Figure 11 As shown, a user profile generation method is provided, including the following steps: obtaining transaction data and video information at the device or account level; generating a user profile at the device or account level based on the transaction data and video information at the device or account level; generating a user profile at the device level based on the user profile at the device or account level and the mapping relationship between devices and accounts; obtaining a user profile through a tag propagation algorithm based on the user profile at the device level and the user's friend relationships; and inputting the user profile into a recommendation system.

[0113] In one embodiment, such as Figure 12As shown, a user profile generation method is provided, including the following steps: obtaining the user's historical viewing behavior and video information (including title, duration and category) on other video platforms; obtaining the user's viewing profile based on the user's historical viewing behavior and video information; inputting the user's viewing profile into the recommendation system of the short video platform, and recommending and displaying corresponding videos to the user through a recall model, a coarse ranking model and a fine ranking model in sequence.

[0114] In the above embodiments, the recommendation system filters information by predicting user ratings and preferences for items. The generated viewing profiles are used in video recommendation services. High-weight tags (those considered to have high confidence) are used for video insertion strategies, while tags with average weight (those considered to have moderate confidence) are used for video recall and ranking. After bucket testing (A / B experiments), it effectively improves metrics such as viewing time, next-day retention (the percentage of users who access the site on a given day and continue to access it on the following day), and video views (vv), demonstrating significant results.

[0115] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0116] In one embodiment, such as Figure 13 As shown, a tagged information generation device 1300 is provided. This device can be a software module, a hardware module, or a combination of both as part of a computer device. Specifically, the device includes: an acquisition module 1310, an extraction module 1320, a determination module 1330, and a generation module 1340, wherein:

[0117] The acquisition module 1310 is used to acquire the target object's historical viewing records on other video platforms when the target object's viewing records on the current video platform do not meet the preset conditions. The historical viewing records include video information of each video watched and corresponding viewing behavior information.

[0118] The extraction module 1320 is used to extract video tags from the video information of each video.

[0119] The determination module 1330 is used to determine the weight of each video tag based on the video information of each video under each video tag and the corresponding viewing behavior information.

[0120] The generation module 1340 is used to generate tagged information of the target object based on each video tag and its weight, and the current video platform recommends videos for the target object based on the tagged information.

[0121] In one embodiment, when acquiring the historical viewing records of a target object on other video platforms, the acquisition module 1310 is specifically used to: determine the account associated with the target object on other video platforms, and acquire the historical viewing records of the account on other video platforms; when it is determined that the account corresponds to a user based on the login device information of the account on other video platforms, the historical viewing records of the account on other video platforms are used as the historical viewing records of the target object on other video platforms; when it is determined that the account corresponds to multiple users based on the login device information of the account on other video platforms, the device identifier of the target object logging into the current video platform is acquired, and the historical viewing records associated with the device identifier are filtered from the historical viewing records of the account on other video platforms and used as the historical viewing records of the target object on other video platforms.

[0122] In one embodiment, the acquisition module 1310 is further configured to: determine that the account corresponds to multiple users when the number of mobile devices that the account has logged into other video platforms within the most recent preset time period exceeds a first threshold.

[0123] In one embodiment, the acquisition module 1310 is further configured to: determine that the account corresponds to multiple users when the number of the same type of mobile devices that the account has logged into other video platforms within the most recent preset time period exceeds a second threshold.

[0124] In one embodiment, the video information includes video category and video title; when the extraction module 1320 extracts the video tags of each video from the video information of each video, it is specifically used to: when the video category of each video is TV series, movie, variety show or animation, use the video title of each video as the video tag of each video.

[0125] In one embodiment, the video information includes the video duration, and the viewing behavior information includes the number of views and the viewing duration and viewing time of each view. When determining the weight of each video tag based on the video information and corresponding viewing behavior information of each video under each video tag, the determining module 1330 is specifically used to: for each video under each video tag, for each view, obtain the viewing factor of the video under this view based on the viewing duration and viewing time of the video under this view; and fuse the viewing factors of each video under each view to obtain the weight of the video tag.

[0126] In one embodiment, when determining the viewing factor of a video based on its viewing duration and viewing time, the determining module 1330 is specifically configured to: obtain a viewing duration factor based on the viewing duration of the video; obtain a completion rate factor based on the ratio of the viewing duration to the video duration; obtain a time decay factor based on the time difference between the viewing time and the current time; and obtain the viewing factor based on the viewing duration factor, completion rate factor, and time decay factor.

[0127] In one embodiment, the acquisition module 1310 is further configured to: when the number of video tags does not meet the quantity requirement, acquire the associated object information of the target object, the associated object information including the relationship closeness between the target object and each associated object; the determination module 1320 is further configured to: obtain the propagation tag and its weight of the target object based on the relationship closeness between the target object and each associated object, and the tags and their weights of each associated object; when the generation module 1340 generates the tagging information of the target object based on each video tag and its weight, it is specifically configured to: generate the tagging information of the target object based on each video tag and its weight, and the propagation tag and its weight of the target object.

[0128] Specific limitations regarding the tagging information generation device can be found in the limitations of the tagging information generation method described above, and will not be repeated here. Each module in the aforementioned tagging information generation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0129] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 14 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a tagged information generation method.

[0130] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 15 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a tagged information generation method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0131] Those skilled in the art will understand that Figure 14 or Figure 15 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0132] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0133] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0134] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0135] It should be understood that the terms "first," "second," etc., used in the above embodiments are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Furthermore, in the description of this application, unless otherwise stated, "a plurality of" means at least two.

[0136] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0137] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0138] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for generating tagged information, characterized in that, The method includes: When the target object's viewing history on the current video platform does not meet the preset conditions, the target object's historical viewing history on other video platforms is obtained. The historical viewing history includes video information of each video watched and corresponding viewing behavior information. The video information includes the video duration, and the viewing behavior information includes the number of times it was watched, as well as the viewing duration and viewing time of each time it was watched. Extract video tags from the video information of each video; For each video under each video tag, for each viewing, obtain the viewing duration factor of the video under this viewing based on the viewing duration of the video in this viewing; The completion rate factor of the video under this viewing is obtained based on the ratio of the viewing time to the video duration under this viewing. The time decay factor of the video under this viewing period is obtained based on the time difference between the viewing time of the video under this viewing period and the current time. The viewing factor of the video under this viewing is obtained based on the viewing duration factor, completion rate factor, and time decay factor of the video under this viewing. The viewing factors of each video under the video tag are fused under each viewing to obtain the weight of the video tag; Based on the video tags and their weights, tagged information of the target object is generated, and the current video platform recommends videos for the target object based on the tagged information.

2. The method according to claim 1, characterized in that, The step of obtaining the target object's historical viewing records on other video platforms includes: Identify the accounts associated with the target object on other video platforms, and obtain the historical viewing records of the accounts on those other video platforms; When it is determined that the account corresponds to a user based on the login device information of the account on other video platforms, the historical viewing records of the account on other video platforms are used as the historical viewing records of the target object on other video platforms. When it is determined that the account corresponds to multiple users based on the login device information of the account on other video platforms, the device identifier of the target object logging into the current video platform is obtained, and the historical viewing records associated with the device identifier are filtered from the historical viewing records of the account on other video platforms as the historical viewing records of the target object on other video platforms.

3. The method according to claim 2, characterized in that, It also includes at least one of the following: When the number of mobile devices that the account has logged into on other video platforms within the most recent preset time period exceeds a first threshold, it is determined that the account corresponds to multiple users; When the number of times the account logs into the same type of mobile device on other video platforms within the recent preset time period exceeds the second threshold, it is determined that the account corresponds to multiple users.

4. The method according to claim 1, characterized in that, The video information includes the video category and video title; Extracting video tags from the video information of each video includes: When the video category of each video is TV series, movie, variety show or animation, the video title of each video will be used as the video tag of each video.

5. The method according to any one of claims 1 to 4, characterized in that, Also includes: When the number of video tags does not meet the requirement, obtain the associated object information of the target object, the associated object information including the relationship between the target object and each associated object; Based on the closeness of the relationship between the target object and each of the associated objects, and the tags and weights of each of the associated objects, the propagation tags and weights of the target object are obtained; The step of generating the tagged information of the target object based on each of the video tags and their weights includes: Based on the video tags and their weights, and the propagation tags and their weights of the target object, the tagged information of the target object is generated.

6. A tagging information generation device, characterized in that, The device includes: The acquisition module is used to acquire the target object's historical viewing records on other video platforms when the target object's viewing records on the current video platform do not meet the preset conditions. The historical viewing records include video information of each video watched and corresponding viewing behavior information. The video information includes the video duration, and the viewing behavior information includes the number of times the video was watched, as well as the viewing duration and viewing time of each video. The extraction module is used to extract video tags from the video information of each video. The determination module is used to, for each video under each video tag, for each viewing, obtain a viewing duration factor for the video in the current viewing based on the viewing duration of the video in the current viewing; obtain a completion rate factor for the video in the current viewing based on the ratio of the viewing duration to the video duration in the current viewing; obtain a time decay factor for the video in the current viewing based on the time difference between the viewing time of the video in the current viewing and the current time; obtain a viewing factor for the video in the current viewing based on the viewing duration factor, completion rate factor, and time decay factor; and fuse the viewing factors of each video under the video tag in each viewing to obtain the weight of the video tag. The generation module is used to generate tagged information of the target object based on each video tag and its weight, and the current video platform recommends videos for the target object based on the tagged information.

7. The tagging information generation device according to claim 6, characterized in that, The acquisition module is further configured to determine the accounts associated with the target object on other video platforms, and acquire the historical viewing records of the accounts on other video platforms; when it is determined that the account corresponds to one user based on the login device information of the account on other video platforms, the historical viewing records of the account on other video platforms are used as the historical viewing records of the target object on other video platforms; when it is determined that the account corresponds to multiple users based on the login device information of the account on other video platforms, the device identifier of the target object logging into the current video platform is acquired, and the historical viewing records associated with the device identifier are filtered from the historical viewing records of the account on other video platforms as the historical viewing records of the target object on other video platforms.

8. The tagging information generation device according to claim 7, characterized in that, The acquisition module is further configured to perform at least one of the following: when the number of mobile devices that the account has logged into on other video platforms within the most recent preset time period exceeds a first threshold, determine that the account corresponds to multiple users; when the number of the same type of mobile devices that the account has logged into on other video platforms within the most recent preset time period exceeds a second threshold, determine that the account corresponds to multiple users.

9. The tagging information generation device according to claim 6, characterized in that, The video information includes video category and video title; the extraction module is also used to use the video title of each video as the video tag of each video when the video category of each video is TV series, movie, variety show or animation.

10. The tagging information generation apparatus according to any one of claims 6 to 9, characterized in that, The acquisition module is further configured to acquire the associated object information of the target object when the number of video tags does not meet the quantity requirement, the associated object information including the relationship closeness between the target object and each associated object; the determination module is further configured to obtain the propagation tag and its weight of the target object based on the relationship closeness between the target object and each associated object, and the tags and their weights of each associated object; the generation module is further configured to generate the tagging information of the target object based on each video tag and its weight, and the propagation tag and its weight of the target object.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 5.

12. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

Images