Method, apparatus, device and medium for generating a shared content list for a team

Abstract

The present application relates to a method for generating a shared content list for a team, including: randomly extracting a content to be calculated from a content database and obtaining the content tags of the content to be calculated; calculating personal preference values based on the personal behavior data of each dimension feature; calculating team preference values based on the team behavior data of each dimension feature; calculating a preference matching value for the content to be calculated according to the personal preference values, the team preference values and the corresponding weight values; and when the preference matching value is greater than a preference matching threshold, adding the content to be calculated to the shared content list. By combining the preference data of individuals and teams, the present application can more accurately predict and recommend content that members may be interested in.

Description

Technical Field

[0001] This application relates to the technical field of content recommendation, and particularly to a method, device, computing device and computer-readable storage medium for generating a shared content list for a team. Background Art

[0002] In modern team collaboration, the construction and management of a shared content list is one of the key factors in improving team collaboration and overall work efficiency. Team members often need to share a large amount of content, such as materials in the form of videos and audios, for learning, communication and cooperation. These shared contents not only contribute to the improvement of members' skills, but also ensure that team members can synchronously obtain the latest material information and maintain the consistency and efficiency of collaboration.

[0003] However, the current shared content recommendation system fails to fully consider the common needs of team members, resulting in the recommended content being overly concentrated in specific directions or categories, or ignoring the interests and needs of other members in the team. This singularity and inaccuracy of content recommendation make the content in the shared content list unable to truly reflect the overall needs of the team and reduce the value of the shared content. Therefore, how to accurately generate a shared content list recognized by all team members remains an urgent problem to be solved. Summary of the Invention

[0004] In view of the technical problems existing in the prior art, this application proposes a method for generating a shared content list for a team, including: randomly extracting a content to be calculated from a content database, and obtaining the content tags of the content to be calculated, where the content tags are used to describe multiple dimensional features to which the content to be calculated belongs; calculating personal preference values according to the personal behavior data of each dimensional feature, where the personal behavior data is used to represent: the interaction behavior data between individual users in the system and the content in the content database, and the personal preference values are used to represent the popularity of the dimensional features to which the content belongs among the individual user group; calculating team preference values according to the team behavior data of each dimensional feature, where the team behavior data is used to represent: the interaction behavior data between all members of the team and the content in the shared content list, and the team preference values are used to represent the popularity of the dimensional features to which the content belongs among team members; calculating a preference matching value of the content to be calculated according to the personal preference values, team preference values and corresponding weight values, where the preference matching value is used to represent the popularity of the content among individual users and team members; when the preference matching value is greater than a preference matching threshold, adding the content to be calculated to the shared content list.

[0005] The method described above for calculating the personal preference value based on the personal behavior data of each dimensional feature includes: counting the personal behavior data of the type to be calculated in each dimensional feature; counting the total sum of the personal behavior data of all types in each dimensional feature; obtaining the personal feature preference value of each dimensional feature according to the ratio of the personal behavior data of the type to be calculated in each dimensional feature to the total sum of the personal behavior data of all types, where the personal feature preference value is used to represent the popularity of the content type in the personal user group; and calculating the geometric mean of the personal feature preference values of all dimensional features to obtain the personal preference value.

[0006] The method described above for obtaining the personal feature preference value of each dimensional feature according to the ratio of the personal behavior data of the type to be calculated in each dimensional feature to the total sum of the personal behavior data of all types includes: obtaining the personal behavior data feature preference value of the i-th type in each dimensional feature according to the ratio of the i-th personal behavior data of the type to be calculated in each dimensional feature to the total sum of the i-th personal behavior data of all types, where i ∈ c and the total number of types of personal behavior data is c; and performing weighted normalization calculation on the c personal behavior data feature preference values to obtain the personal feature preference value of each dimensional feature.

[0007] The method described above for calculating the team preference value based on the team behavior data of each dimensional feature includes: counting the team behavior data of the type to be calculated in each dimensional feature; counting the total sum of the team behavior data of all types in each dimensional feature; obtaining the team feature preference value of each dimensional feature according to the ratio of the team behavior data of the type to be calculated in each dimensional feature to the total sum of the team behavior data of all types, where the team feature preference value is used to represent the popularity of the content type among team members; and calculating the geometric mean of the team feature preference values of all dimensional features to obtain the team preference value.

[0008] The method described above for obtaining the team feature preference value of each dimensional feature according to the ratio of the team behavior data of the type to be calculated in each dimensional feature to the total sum of the team behavior data of all types includes: obtaining the team behavior data feature preference value of the I-th type in each dimensional feature according to the ratio of the I-th team behavior data of the type to be calculated in each dimensional feature to the total sum of the I-th team behavior data of all types, where I ∈ C and the total number of types of team behavior data is C; and performing weighted normalization calculation on the C team behavior data feature preference values to obtain the team feature preference value of each dimensional feature.

[0009] The method described above calculates the team preference value based on the team behavior data of each dimensional feature and the behavior data of team members on the shared content list.

[0010] The method described above calculates the team preference value using the following formula:

[0011]

[0012] where score_g is the team preference value, B k is the team behavior data feature preference value of the k-th dimension feature. There are a total of z dimension features, α I is the weight value of the i-th type of team behavior data. There are a total of s types of team behavior data, X is the number of shared content lists, sum(v Ix ) is the i-th type of team behavior data to which the content to be calculated in the x-th shared content list belongs, and sum(v_all Ix ) is the total sum of the i-th type of team behavior data of all types in the x-th shared content list. is the weighting coefficient of the shared content list x, and the weighting coefficient is calculated based on the behavior data of team members for the shared content list.

[0013] In the method described above, the behavior data of team members for the shared content list includes: one or more of the click data of team members for the shared content list, the sharing quantity, and whether to delete the shared content list.

[0014] In the method described above, the personal preference value is calculated based on the personal behavior data of each dimension feature within multiple specified time periods; and / or the team preference value is calculated based on the team behavior data of each dimension feature and the behavior data of team members for the shared content list within multiple specified time periods.

[0015] In the method described above, the individual users in the system include: any one of all registered users in the system, users who have joined a team among all registered users in the system, and users who have joined the target team among all registered users in the system.

[0016] In the method described above, the content includes one or more of dance videos and dance music, and the dance music is the song separated from the dance video.

[0017] In the method described above, when the content to be calculated is a dance video or dance music, the dimension features of the dance video or dance music are decomposed into: one or more of dance type features, movement type features, and rhythm type features.

[0018] According to another aspect of the present application, a shared content list generation device for a team is proposed, including: an extraction module for randomly extracting a content to be calculated from a content database and obtaining content tags of the content to be calculated, where the content tags are used to describe multiple dimensional features to which the content to be calculated belongs; a first calculation module for calculating a personal preference value according to personal behavior data of each dimensional feature, where the personal behavior data is used to represent: interaction behavior data between individual users in the system and the content in the content database, and the personal preference value is used to represent the popularity of the content among the individual user group; a second calculation module for calculating a team preference value according to team behavior data of each dimensional feature, where the team behavior data is used to represent: interaction behavior data between all members in the team and the content in the shared content list, and the team preference value is used to represent the popularity of the content among team members; a third calculation module for calculating a preference matching value of the content to be calculated according to the personal preference value, the team preference value and the corresponding weight value, where the preference matching value is used to represent the popularity of the content among individual users and team members; a judgment module for adding the content to be calculated to the shared content list when the preference matching value is greater than a preference matching threshold.

[0019] According to another aspect of the present application, a computing device is proposed, including: a processor; a memory storing a computer program, which when executed by the processor, implements the above-mentioned method for generating a shared content list for a team.

[0020] According to another aspect of the present application, a computer-readable storage medium is proposed, where computer instructions are stored in the computer-readable storage medium, and when the computer instructions are executed by a processor, the above-mentioned method for generating a shared content list for a team is implemented.

[0021] When the present application determines whether to add the content to be calculated to the shared content list, it refers to the personal user behavior data and timely adds popular content to the shared content list, avoiding simple "one-size-fits-all" recommendations and making the content type too single. Further, when determining whether to add the content to be calculated to the shared content list, it also considers the behavior data of each member in the team, making the content in the shared content list conform to the common hobbies of team members, which is conducive to promoting interaction and cooperation within the team. By combining personal and team preference data, the present application can more accurately predict and recommend content that members may be interested in. Description of the Drawings

[0022] Next, the preferred embodiments of the present application will be further described in detail with reference to the drawings, where:

[0023] Figure 1Flowchart of a method for generating a shared content list for a team according to an embodiment of the present application.

[0024] Figure 2 Flowchart of a method for calculating personal preference values in step S120.

[0025] Figure 3 Flowchart of a method for calculating team preference values in step S130.

[0026] Figure 4 Interface diagram of a shared dance video list and a shared dance music list according to an embodiment of the present application.

[0027] Figure 5 Device for generating a shared content list for a team according to an embodiment of the present application.

[0028] Figure 6 Block diagram of the structural principle of a computing device according to an embodiment of the present application. Detailed implementation

[0029] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the scope of protection of the present application.

[0030] In the following detailed description, reference may be made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific embodiments in which the application may be practiced. In the drawings, similar reference numerals generally refer to similar components in different figures. The various specific embodiments of the present application are described in sufficient detail below to enable those of ordinary skill in the art with relevant knowledge and technology to implement the technical solutions of the present application. It should be understood that other embodiments may be utilized or structural, logical or electrical changes may be made to the embodiments of the present application.

[0031] The shared content list of the present application is a list containing various shareable and accessible content. In the context of teamwork or project management, it is usually used to indicate materials, documents, videos, audio or other forms of content that all team members can view, download or use. For example, in a dance team, there may be a "shared content list" listing all dance videos, dance music or rehearsal materials, and team members can access and use these contents as needed. The shared content list can be in digital form, such as stored in the resource library of an application or a database on a cloud server.

[0032] When the team is a dance team, the content to be calculated in this application includes one or more of a dance video and dance music. A dance video is an audio-visual file recording a dance demonstration, usually containing dance movements and the performance of dancers. These videos may be used as rehearsal references, teaching materials, or performance replays.

[0033] Dance music refers to the music that matches the dance performed in the dance video. The dance music can be a song "separated" from the dance video. That is to say, the original audio track in the dance video contains the music. Extract or separately obtain these audio tracks (i.e., dance music) from the video and add them to the shared content list for use in practice, performance, or other related activities.

[0034] Figure 1 It is a flowchart of a method for generating a shared content list for a team according to an embodiment of this application. As Figure 1 shown, the method includes:

[0035] Step S110, randomly select a content to be calculated in the content database, and obtain the content label of the content to be calculated. The content label is used to describe multiple dimensional features to which the content to be calculated belongs;

[0036] Step S120, calculate the personal preference value according to the personal behavior data of each dimensional feature. The personal behavior data is used to represent: the interaction behavior data between individual users in the system and the content in the content database. The personal preference value is used to represent the popularity of the dimensional feature to which the content belongs among the individual user group;

[0037] Step S130, calculate the team preference value according to the team behavior data of each dimensional feature. The team behavior data is used to represent: the interaction behavior data between all members of the team and the content in the shared content list. The team preference value is used to represent the popularity of the dimensional feature to which the content belongs among the team members;

[0038] Step S140, calculate the preference matching value of the content to be calculated according to the personal preference value, the team preference value, and the corresponding weight value. The preference matching value is used to represent the popularity of the content among individual users and team members;

[0039] Step S150, when the preference matching value is greater than the preference matching threshold, add the content to be calculated to the shared content list.

[0040] In step S110, a content to be calculated is randomly selected from the content database. The content database may contain different types of content, such as dance videos, dance music, dance tutorials, etc. Each content item has one or more content tags used to describe its characteristics in multiple dimensions, such as the type, style, difficulty, duration, music type, etc. of the content. For example, when the content to be calculated is a dance video, the content tags of a certain video may include multiple dimensional characteristics such as "hip-hop", "basic movements", "3 minutes", and "dynamic music".

[0041] In step S120, personal behavior data of the content to be calculated in multiple dimensional characteristics is obtained and a personal preference value is calculated. According to the previous calculation method, it is directly calculated based on the personal behavior data of the content to be calculated. Different from this, this application does not focus on whether the content is liked by the personal user group, but on whether the content tags represented by the content (that is, the type to which the content belongs) are liked by the personal user group, so as to obtain the content types liked by the personal user group. Interaction behavior data is the data statistically obtained from the user's operations on the content, such as one or more of the number of plays, play duration, number of likes, number of collections, and number of shares of the content. In this application, the personal user can be any one of the following three groups of people:

[0042] ① All registered users in the system: This group of people includes users who have not joined any team and users who have joined certain teams. The system will analyze the behavior data of all users to obtain the overall user preference information in the system;

[0043] ② Registered users in the system who have joined a team: This group of people refers to all users who have registered in the system and have joined at least one team. The behavior data of this type of users can reflect their preferences in different team or collective environments;

[0044] ③ Registered users in the system who have joined the target team: This group of people specifically refers to users who have registered and joined the target team. For example, the target team users refer to the team corresponding to the shared content list, and the behavior data of these users will be used to calculate the personal preference values of the members in the team.

[0045] In step S130, team behavior data of the content to be calculated in multiple dimensional characteristics is obtained and a team preference value is calculated. The calculation method of the team preference value is similar to that of the personal preference value, except that it is based on the interaction behavior data of all team members, rather than the data of a single user.

[0046] In step S140, by combining the personal preference value and the team preference value and according to the set weight value, the preference matching value of the content to be calculated is calculated. The weight value is used to measure the importance of personal preference and team preference in the final decision. In some embodiments, the weight value of the team preference value is greater than the weight value of the personal preference value to focus on whether the content to be calculated is liked by team members.

[0047] In step S150, it is judged whether to add the content to be calculated to the shared content list according to a preset preference matching threshold. If the preference matching value of the content to be calculated is greater than the threshold, the content is added to the shared content list for team members to access and use.

[0048] In the solution of the present application, when judging whether to add the content to be calculated to the shared content list, personal user behavior data is referred to. The behavior data of each user reflects their interaction with the content in the system. These data can identify popular content types, styles or tags, and add popular content to the shared content list in a timely manner, avoiding a simple "one-size-fits-all" recommendation and making the content type too single.

[0049] Furthermore, when judging whether to add the content to be calculated to the shared content list, the behavior data (such as likes, comments, shares, etc.) of each member in the team is also considered, so that the content in the shared content list conforms to the common hobbies of team members, making the popularity of the content higher and facilitating interaction and cooperation within the team. By combining personal and team preference data, the present application can more accurately predict and recommend content that members may be interested in. Compared with the traditional single recommendation mechanism, the recommendation mechanism that combines team and personal data can significantly improve the accuracy and diversity of content recommendation, making team members more satisfied with the recommended content.

[0050] Figure 2 It is a flowchart of the method for calculating the personal preference value in step S120. As Figure 2 shown, the method includes:

[0051] Step S121, statistically analyzing the personal behavior data of the type to which the content to be calculated belongs in each dimension feature;

[0052] Step S122, statistically analyzing the total sum of personal behavior data of all types in each dimension feature;

[0053] Step S123, obtaining the personal feature preference value of each dimension feature according to the ratio of the personal behavior data of the type to which the content to be calculated belongs and the total sum of personal behavior data of all types in each dimension feature. The personal feature preference value is used to represent the popularity of the content type among the personal user group;

[0054] Step S124: Calculate the geometric mean of the personal characteristic preference values of all dimensional characteristics to obtain the personal preference value.

[0055] In step S121, the dimensional characteristic refers to each descriptive characteristic of the content. Different content types under each dimensional characteristic will contain certain interaction data (such as clicks, shares, comments, etc.), and these behavioral data reflect the degree of interest of individual users in the content type.

[0056] When the content to be calculated is a dance video or dance music, the dimensional characteristics of the dance video or dance music are decomposed into one or more of the following: dance type characteristic, dance content characteristic, movement type characteristic, and rhythm type characteristic. Among them, the dance type characteristic includes: classical dance, square dance, ethnic dance, pop dance; the dance content characteristic includes one or more of the following: teaching content, performance content, competition content, and daily practice. The movement type characteristic includes: movement difficulty characteristic (beginner, easy, moderate, slightly difficult...), movement intensity characteristic (low intensity, medium intensity, high intensity...), movement part characteristic (shoulder and neck, waist and abdomen, knees, whole body...); the rhythm type characteristic includes: rhythm speed characteristic (slow, medium, fast...), rhythm emotion characteristic (soothing, lively, high-pitched...).

[0057] For example, if the content to be calculated is dance video V, the dance type characteristic of this video is: square dance; the movement difficulty characteristic is: moderate; the movement intensity characteristic is: low intensity. Then, count all the personal user interaction data of the dance type "square dance"; count all the personal user interaction data of the movement difficulty characteristic "moderate"; count all the personal user interaction data of the movement intensity characteristic "low intensity".

[0058] In step S122, calculate the total sum of the personal behavior data of all types in each dimensional characteristic to provide a basis for subsequent calculations. As mentioned above, still taking the above dance video V as an example, count all the personal user interaction data of all dance types; count all the personal user interaction data of all movement difficulty characteristics; count all the personal user interaction data of all movement intensity characteristics.

[0059] In step S123, the personal characteristic preference value is represented by the ratio of the personal behavior data of the type to which the content to be calculated belongs in each dimensional characteristic to the total sum of the personal behavior data of all types. Suppose the playback volume of the user watching the "square dance" video is 300 times, and among all dance types (classical dance, square dance, ethnic dance, pop dance), the user has watched a total of 6000 times. Then the personal characteristic preference value = 300 / 6000 = 0.05, and the personal characteristic preference value is 0.05.

[0060] In step S124, the personal preference value is calculated by the geometric mean according to the personal characteristic preference values of each dimension feature. The geometric mean is used to comprehensively consider the influence of multiple dimension features on personal preferences and avoid the excessive influence of a single dimension feature on the final preference value. The above method can accurately model the user's interests, provide strong support for content recommendation, and ensure that the recommended content meets the multi-dimensional preferences of individual users.

[0061] As can be seen from the above method, the personal behavior data includes various behavior data, such as: the number of plays, the play duration, the number of likes, the number of collections, and the number of shares, etc. For each type of behavior data, a personal behavior data characteristic preference value can be calculated.

[0062] According to an embodiment of the present application, obtaining the personal characteristic preference value of each dimension feature according to the ratio of the personal behavior data of the type to be calculated in each dimension feature to the total sum of the personal behavior data of all types includes:

[0063] Obtaining the personal behavior data characteristic preference value of the i-th type in each dimension feature according to the ratio of the i-th type of personal behavior data of the type to be calculated in each dimension feature to the total sum of the i-th type of personal behavior data of all types, where i ∈ c, and the total number of types of personal behavior data is c;

[0064] Performing weighted normalization calculation on the c personal behavior data characteristic preference values to obtain the personal characteristic preference value of each dimension feature.

[0065] The above method refines the personal behavior data of the type to be calculated in each dimension feature, and calculates the preference value of each behavior data characteristic according to different types of behavior data (such as the number of plays, the play duration, the number of likes, etc.). Finally, the personal characteristic preference value of each dimension feature is obtained by weighted normalization calculation. This method can more carefully capture the user's behavior preferences and further optimize content recommendation.

[0066] According to an embodiment of the present application, the personal preference value is calculated using the following formula:

[0067]

[0068] where score_p is the personal preference value, A k is the personal characteristic preference value of the k-th dimension feature, there are a total of z dimension features, sum(v i ) is the total sum of the i-th type of personal behavior data of the type to be calculated, sum(v_all i ) is the total sum of the i-th type of personal behavior data of all types, α i$w_{i}$ is the weight of the $i$-th type of personal behavior data of the type to be calculated, and there are a total of $c$ types of personal behavior data.

[0069] Figure 3 It is the flowchart of the method for calculating the team preference value in step S130. As Figure 3 shown, the method includes:

[0070] Step S131, counting the team behavior data of the type to be calculated in each dimension feature;

[0071] Step S132, counting the total sum of team behavior data of all types in each dimension feature;

[0072] Step S133, obtaining the team feature preference value of each dimension feature according to the ratio of the team behavior data of the type to be calculated in each dimension feature to the total sum of team behavior data of all types, and the team feature preference value is used to represent the popularity of the content type among team members;

[0073] Step S134, calculating the geometric mean of the team feature preference values of all dimension features to obtain the team preference value.

[0074] This method is similar to the method for calculating the personal preference value in Figure 2 , and the only difference is the data statistics. Regarding the specific steps of the method, they will not be elaborated here. The purpose of this method is to calculate the team feature preference value of each dimension feature by statistically analyzing the interaction behavior data of team members, and finally obtain a comprehensive team preference value, which is used to better recommend content that meets the team's interests and improve the participation and satisfaction of team members.

[0075] Furthermore, obtaining the team feature preference value of each dimension feature according to the ratio of the team behavior data of the type to be calculated in each dimension feature to the total sum of team behavior data of all types includes:

[0076] Obtaining the feature preference value of the $I$-th type of team behavior data in each dimension feature according to the ratio of the $I$-th type of team behavior data of the type to be calculated in each dimension feature to the total sum of the $I$-th type of team behavior data of all types, where $I\in C$, and the total number of types of team behavior data is $C$;

[0077] Performing weighted normalization calculation on the $C$ types of team behavior data feature preference values to obtain the team feature preference value of each dimension feature.

[0078] According to an embodiment of the present application, the team preference value is calculated using the following formula:

[0079]

[0080] Among them, score_g is the team preference value, and B k is the team feature preference value of the k-th dimensional feature. There are a total of z dimensional features. sum(v I ) is the sum of the I-th type of team behavior data of the content type to be calculated. sum(v_all I ) is the sum of the I-th type of team behavior data of all types. α I is the weight value of the I-th type of team behavior data of the content type to be calculated. There are a total of S types of team behavior data.

[0081] According to another embodiment of the present application, the team preference value is calculated based on the team behavior data of each dimensional feature and the behavior data of team members on the shared content list.

[0082] In the above embodiment, the team behavior data is the behavior data generated by team members taking the content in the shared content list as the object (such as the number of plays, the number of collections, the number of shares, etc.), and the behavior data of team members on the shared content list is the behavior data generated by team members taking the shared content list as the object. The behavior data of team members on the shared content list includes: the click data of team members on the shared content list, the number of shares, and one or more of whether to delete the content in the shared content list. By analyzing the behavior data of team members on the shared content list, the interests and participation degrees of team members in various types of content in the shared content list can be understood, and thus the overall preference of the team can be inferred.

[0083] In the above embodiment, through the statistics and analysis of the interaction behavior data between team members and the shared content list, considering the interaction behavior of team members with the content and the interaction behavior of the team with the shared content list, a team preference value that can reflect the overall preference of the team is finally obtained. This preference value can be used to determine whether the content meets their interests and improve the recommendation effect of the content.

[0084] According to another embodiment of the present application, the team preference value is calculated using the following formula:

[0085]

[0086] Among them, score_g is the team preference value, and B k is the team behavior data feature preference value of the k-th dimensional feature. There are a total of z dimensional features. α I is the weight value of the I-th type of team behavior data. There are a total of S types of team behavior data. X is the number of shared content lists. sum(v Ix ) is the I-th type of team behavior data of the content type to be calculated in the x-th shared content list. sum(v_all Ix) is the total sum of the I-th type of team behavior data for all types in the x-th shared content list, is the weighting coefficient of the shared content list x, and the weighting coefficient is calculated based on the behavior data of team members for the shared content list.

[0087] As can be seen from the above formula, when the team has multiple shared content lists, multiple weighting coefficients are calculated respectively according to the behavior data of each team member for each shared content list, and then the team preference value is calculated.

[0088] According to an embodiment of the present application, the weighting coefficient is calculated using the following formula:

[0089]

[0090] When the shared content list is in normal use, the weighting coefficient β is calculated using formula (1), and when the shared content list is deleted, the weighting coefficient β is calculated using formula (2), where 、 、e and f are weight values, sum(E) is the total click volume of the shared content list, and sum(F) is the total share volume of the shared content list.

[0091] According to an embodiment of the present application, the personal preference value is calculated based on the personal behavior data of each dimension feature within multiple specified time periods; and / or

[0092] The team preference value is calculated based on the team behavior data of each dimension feature and the behavior data of team members for the shared content list within multiple specified time periods.

[0093] In the above embodiment, by dividing the user behavior data of individuals and teams into multiple time periods in chronological order from far to near, the interest changes of the personal and / or team preference values in different time periods can be obtained, and the preference trends of users and teams can be captured more accurately for more personalized content recommendation.

[0094] According to an embodiment of the present application, the preference matching value is calculated using the following formula:

[0095]

[0096] Among them, S is the preference matching value, μ1 is the weight value of the personal preference value, μ2 is the weight value of the team preference value, δ ti is the weight value of the personal preference value in the ti time period and and , the time period is closer to the current time than the time period, and there are a total of m time periods, is the weight value of the team preference value for the tj time period and and , the time period is closer to the current time than the time period. There are a total of n time periods, and score_p ti is the personal preference value for the ti time period, and score_g tj is the team preference value for the tj time period.

[0097] The above describes the implementation manner and advantages brought by the embodiments of the present application through multiple embodiments. The following takes the example of a dance video to describe in detail the specific processing process of the embodiments of the present application.

[0098] Using the above method, filter dance videos in the dance video database of the server and add them to the shared video list of Dance Team A (hereinafter referred to as Team A) and generate a dance music list. There are 10 dance members in Team A. The summary of the historical data of the team members is as follows in Table 1: (Note: The time is divided from far to near into the recent 3 years, the recent half year, and the recent 1 month. Among them, the recent 3 years do not include the recent half year, and the recent half year does not include the recent 1 month.)

[0099] Table 1

[0100]

[0101] It can be seen from Table 1 that the individual users in the above system are the users who have joined Team A among all the registered users in the system. Decompose the dimensional features of the dance videos in the content database into dance category features and action difficulty features. The user's interaction behavior data for the content includes: the number of play times and the number of favorite times, and the user's behavior data for the dance music list includes the number of click times and the number of share times.

[0102] Combined with Table 1, in the formula of this method, the parameters and assigned weights are as shown in Table 2 below:

[0103] Table 2

[0104]

[0105] First, randomly select a dance video D from the content database. The dance category feature of this video is square dance, and the action difficulty feature is moderate. Then calculate in combination with the data in Tables 1-2 and the above formula.

[0106] (1) Personal characteristic preference value:

[0107] Personal characteristic preference value A1-1 for square dance in the recent 3 years:

[0108]

[0109] Personal characteristic preference value A2-1 with moderate movement difficulty in the past 3 years:

[0110]

[0111] The personal preference value of dance video D in the past 3 years is:

[0112]

[0113] Similarly, the personal characteristic preference value A2-1 of square dance in the past half year = 0.59;

[0114] The personal characteristic preference value A2-2 with moderate movement difficulty in the past half year = 0.68;

[0115] The personal preference value of dance video D in the past half year is: score_p2 = 0.63.

[0116] Similarly, the personal characteristic preference value A3-1 of square dance in the past 1 month = 0.58;

[0117] The personal characteristic preference value A3-2 with moderate movement difficulty in the past 1 month = 0.81;

[0118] The personal preference value of dance video D in the past 1 month is: score_p3 = 0.69.

[0119] (2)Team characteristic preference value

[0120] From the data in Table 1 and Table 2, the weighted coefficients β of the two song lists of dance team A are as shown in Table 3 below:

[0121] Table 3

[0122]

[0123] Combined with Table 1-3 and the above formula, it is calculated that:

[0124] The team characteristic preference value B1-1 of square dance in the past 3 years:

[0125]

[0126] The team characteristic preference value B2-1 with moderate movement difficulty in the past 3 years:

[0127]

[0128] The team preference value of dance video D in the past 3 years is:

[0129]

[0130] Similarly, the team characteristic preference value B2-1 of square dance in the past half year = 0.79;

[0131] The preference value B2-2 of the team with moderate action difficulty in the past six months is 0.61;

[0132] The team preference value of the dance video D in the past six months is: score_ g2 = 0.69.

[0133] Similarly, the preference value B3-1 of the square dance team in the past month is 0.74;

[0134] The preference value B3-2 of the team with moderate action difficulty in the past month is 0.22;

[0135] The team preference value of the dance video D in the past month is: score_ g3 = 0.40.

[0136] (3) Preference matching value

[0137]

[0138] Then the preference matching value of the dance video D is 0.55. Assuming that the preference matching threshold is 0.4, the dance video D will be added to the shared video list until the number of dance videos in the shared video list reaches the preset quantity threshold (such as 20), and the calculation will stop.

[0139] The shared video list is displayed on the client of the dance team members. For example, an entrance to the shared dance video list is displayed in a specific area of the dance team space page. Clicking on the entrance can enter the list details, which include but are not limited to information such as the video cover, video title, and publisher nickname sorted in sequence.

[0140] Moreover, the dance team members can perform online operations and learning on the shared dance video list on the client. For example, the dance team members (including the team leader) can perform online operations on the video list on the function page of the client, such as clicking on the video to watch, removing disliked content, adding other liked content, and adjusting the order of the videos. When adding other liked content, it supports the user to locate the required content through search, or a pop-up window of alternative content can be provided to enable the user to select the required content from the pop-up window.

[0141] Furthermore, the team leader can lock or unlock the list adjusted by the dance team user members on the function page of the client. After locking, the dance team users cannot change it, and after unlocking, it can be changed again. The team leader can perform a refresh operation on the current shared dance video list on the function page of the client, and at this time, the above method is executed again to update the videos in the shared dance video list.

[0142] Finally, a shared dance music list is generated according to the locked dance video list. Such asFigure 4 The interface diagrams of the shared dance video list and the shared dance music list are shown. Among them, the dance music in the shared dance music list corresponds one-to-one with the dance videos in the shared content list. When saving the dance music list, a shared dance music list name is randomly generated and renaming is supported. Among them, the shared video list is used for online learning by dance team users, and the shared dance music list is used for offline speaker playback by the dance team. The videos in the dance video list provide content that all members of the dance team users recognize, facilitating online preview of the content to be learned by dance team members, and at the same time facilitating online content selection, offline teaching, and offline following the dance music list by the dance team captain.

[0143] Figure 5 It is a shared content list generation device for a team according to an embodiment of the present application. The device 100 includes:

[0144] An extraction module 110, configured to randomly extract a content to be calculated from a content database and obtain a content tag of the content to be calculated, where the content tag is used to describe multiple dimensional features to which the content to be calculated belongs;

[0145] A first calculation module 120, configured to calculate a personal preference value according to the personal behavior data of each dimensional feature, where the personal behavior data is used to represent: the interaction behavior data between individual users in the system and the content in the content database, and the personal preference value is used to represent the popularity of the content among individual user groups;

[0146] A second calculation module 130, configured to calculate a team preference value according to the team behavior data of each dimensional feature, where the team behavior data is used to represent: the interaction behavior data between all members of the team and the content in the shared content list, and the team preference value is used to represent the popularity of the content among team members;

[0147] A third calculation module 140, configured to calculate a preference matching value of the content to be calculated according to the personal preference value, the team preference value, and the corresponding weight value, where the preference matching value is used to represent the popularity of the content among individual users and team members;

[0148] A judgment module 150, configured to add the content to be calculated to the shared content list when the preference matching value is greater than the preference matching threshold.

[0149] On the other hand, the present application also provides a computing device. Refer to Figure 6 , Figure 6 is the structural principle block diagram of a computing device according to an embodiment of the present application. As Figure 6As shown, the computing device includes a processor 601 and a memory 602 storing computer program instructions; when the processor 601 executes the computer program instructions, the method for generating a shared content list for a team in the foregoing embodiments is implemented.

[0150] Specifically, the processor 601 may include a central processing unit (CPU) or a graphics processing unit (GPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. The memory 602 may include a memory for data or instructions. For example, the memory 602 may be at least one of the following: a hard disk drive (HDD), a read-only memory (ROM), a random access memory (RAM), a floppy disk drive, a flash memory, an optical disc, a magneto-optical disc, a magnetic tape, a universal serial bus (USB) drive, or other physical / tangible memory storage devices. Also, for example, the memory 602 includes a removable or non-removable (or fixed) medium. Again, for example, the memory 602 may be inside or outside the integrated gateway disaster recovery device. The memory 602 may be a non-volatile solid state memory. In other words, generally the memory 602 includes a tangible (non-transitory) computer-readable storage medium (such as a memory device) encoded with executable instructions, and when the executable instructions stored therein are executed by the processor 601 (such as by one or more processors), the method for generating a shared content list for a team in the embodiments of the present application can be implemented.

[0151] In one example, Figure 6 The electronic device shown may further include a communication interface 603 and a bus 610. Among them, the processor 601, the memory 602, and the communication interface 603 are connected through the bus 610 and communicate with each other. The communication interface 603 is mainly used to implement communication between various modules, devices, units, and / or devices in the electronic device.

[0152] The bus 610 includes hardware, software, or both, and can couple the components of the online data flow metering device to each other. For example, the bus can include at least one of the following: Accelerated Graphics Port (AGP) or other graphics buses, Extended Industry Standard Architecture (EISA) bus, Front Side Bus (FSB), HyperTransport (HT) interconnect, Industry Standard Architecture (ISA) bus, InfiniBand interconnect, Low Pin Count (LPC) bus, Memory bus, MicroChannel Architecture (MCA) bus, Peripheral Component Interconnect (PCI) bus, PCI-Express (PCI-X) bus, Serial Advanced Technology Attachment (SATA) bus, Video Electronics Standards Association Local (VLB) bus, or other suitable buses. The bus 610 can include one or more buses. Although the embodiments of the present application describe or illustrate specific buses, the embodiments of the present application can contemplate any suitable bus or interconnect method.

[0153] In another aspect, the embodiments of the present application further provide a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the foregoing method for generating a shared content list for a team is implemented.

[0154] The flowcharts and / or block diagrams of the methods and systems of the embodiments of the present application are described above by way of example, and the relevant aspects are described. It should be understood that each block in the flowchart and / or block diagram, or a combination thereof, can be implemented by computer program instructions, can be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions. For example, these computer program instructions can be provided to a processor of a general-purpose computer, a dedicated computer, or other programmable data processing device to form a machine such that these instructions executed by such a processor enable the implementation of the specified function / action in each block or a combination thereof in the flowchart and / or block diagram. Such a processor can be a general-purpose processor, a dedicated processor, a special application processor, or a field-programmable logic circuit.

[0155] The functional blocks shown in the structural block diagrams of the embodiments of the present application can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a functional card, etc.; when implemented in software, it is a program or a code segment for performing the required tasks. The program or code segment can be stored in a memory, or transmitted via a data signal carried in a carrier wave on a transmission medium or a communication link. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0156] The above embodiments are only for illustrating the present application and are not intended to limit the present application. Those of ordinary skill in the relevant technical field can also make various changes and modifications without departing from the scope of the present application. Therefore, all equivalent technical solutions should also fall within the scope of the disclosure of the present application.

Claims

1. A method for generating a shared content list for a team, characterized in that: include: Randomly extract a content to be calculated from the content database, and obtain a content tag of the content to be calculated, where the content tag is used to describe multiple dimensional features of the content to be calculated; A personal preference value is calculated based on personal behavior data of each dimensional feature, wherein the personal behavior data is used to represent: interaction behavior data between individual users in the system and content in the content database, and the personal preference value is used to represent the popularity of the dimensional feature to which the content belongs in the individual user group; The team preference value is calculated based on the team behavior data of each dimension feature and the behavior data of the team members on the shared content list, wherein the team behavior data is used to represent: the interaction behavior data of all members in the team and the content in the shared content list, and the team preference value is used to represent the popularity of the dimension feature to which the content belongs among the team members, and the behavior data of the team members on the shared content list includes: the click data of the team members on the shared content list, the number of shares, and whether to delete one or more of the shared content list; The preference matching value of the content to be calculated is calculated according to the individual preference value, the team preference value and the corresponding weight value, and the preference matching value is used to indicate the popularity of the content among individual users and team members; When the preference matching value is greater than the preference matching threshold, the content to be calculated is added to the shared content list. The shared content list is stored in the database of the cloud server. It is a list of content for team members to share or access. Team members can perform online operations and study on the shared content list.

2. The method according to claim 1, characterized in that The personal preference values ​​calculated based on the personal behavior data of each dimension feature include: Count the personal behavior data of the type of content to be calculated in each dimension feature; Count the sum of all types of personal behavior data in each dimension feature; Obtain a personal feature preference value for each dimensional feature according to the ratio of the personal behavior data of the type of content to be calculated in each dimensional feature to the sum of personal behavior data of all types, wherein the personal feature preference value is used to indicate the popularity of the type of content in the individual user group; The personal preference value is obtained by calculating the geometric mean of the personal characteristic preference values ​​of all dimensional features.

3. The method according to claim 2, characterized in that The personal characteristic preference value of each dimensional feature is obtained based on the ratio of the personal behavior data of the type of content to be calculated in each dimensional feature to the sum of all types of personal behavior data, including: Obtain the preference value of the i-th personal behavior data feature in each dimensional feature according to the ratio of the i-th personal behavior data of the type to be calculated content in each dimensional feature to the sum of the i-th personal behavior data of all types, where i∈c, and the number of types of personal behavior data is c in total; The c types of personal behavior data feature preference values ​​are weighted and normalized to obtain the personal feature preference value of each dimension feature.

4. The method according to claim 1, characterized in that: The team preference values ​​calculated based on the team behavior data of each dimension feature and the behavior data of team members on the shared content list include: Count the team behavior data of the type of content to be calculated in each dimension feature; Count the total of all types of team behavior data in each dimension feature; Obtain a team feature preference value for each dimensional feature according to the ratio of the team behavior data of the type of content to be calculated in each dimensional feature to the total team behavior data of all types, wherein the team feature preference value is used to indicate the popularity of the type of content among team members; The team preference value is obtained by calculating the geometric mean of the team characteristic preference values ​​of all dimensional characteristics.

5. The method according to claim 4, characterized in that The team feature preference values ​​of each dimensional feature are obtained based on the ratio of the team behavior data of the type of content to be calculated in each dimensional feature to the total team behavior data of all types, including: Obtain the feature preference value of the first type of team behavior data in each dimensional feature according to the ratio of the first type of team behavior data of the type to be calculated in each dimensional feature to the sum of the first type of team behavior data of all types, where I∈C, and the number of types of team behavior data is C in total; The characteristic preference values ​​of C types of team behavior data are weighted and normalized to obtain the team characteristic preference value of each dimension feature.

6. The method according to claim 4, characterized in that The team preference value is calculated using the following formula: Among them, score_g is the team preference value, B k is the team behavior data feature preference value of the kth dimension feature. There are a total of z dimension features. α I is the weight value of the I-th type of team behavior data, there are s types of team behavior data, X is the number of shared content lists, sum(v Ix ) is the I-th type of team behavior data of the type of content to be calculated in the x-th shared content list, sum(v_all Ix ) is the sum of the first type of team behavior data of all types in the xth shared content list, β x is a weighting coefficient of the shared content list x, and the weighting coefficient is calculated according to the behavior data of the team members on the shared content list.

7. The method according to claim 1, characterized in that Calculate the personal preference value based on the personal behavior data of each dimension feature within multiple specified time periods; and / or The team preference value is calculated based on the team behavior data of each dimension feature in multiple specified time periods and the behavior data of team members on the shared content list.

8. The method according to claim 1, characterized in that The individual users in the system include: all users registered in the system, users who have joined a team among all users registered in the system, and any one of users who have joined a target team among all users registered in the system.

9. The method according to any one of claims 1 to 8, characterized in that: The content includes one or more of a dance video and dance music, wherein the dance music is a song separated from the dance video.

10. The method according to claim 1, characterized in that When the content to be calculated is a dance video or dance music, the dimensional features of the dance video or dance music are decomposed into one or more of dance type features, action type features and rhythm type features.

11. A device for generating a shared content list for a team, characterized in that: include: An extraction module, used to randomly extract a content to be calculated from the content database and obtain a content tag of the content to be calculated, wherein the content tag is used to describe multiple dimensional features of the content to be calculated; A first calculation module is used to calculate a personal preference value based on personal behavior data of each dimension feature, wherein the personal behavior data is used to represent: interaction behavior data between a personal user in the system and content in a content database, and the personal preference value is used to represent the popularity of the content in a personal user group; A second calculation module is used to calculate a team preference value based on the team behavior data of each dimension feature and the behavior data of the team members on the shared content list, wherein the team behavior data is used to represent: the interaction behavior data of all members in the team and the content in the shared content list, and the team preference value is used to represent the popularity of the content among the team members, and the behavior data of the team members on the shared content list includes: the click data of the team members on the shared content list, the number of shares, and whether to delete one or more of the shared content list; A third calculation module is used to calculate the preference matching value of the content to be calculated according to the personal preference value and the team preference value and the corresponding weight value, wherein the preference matching value is used to indicate the popularity of the content among individual users and team members; The judgment module is used to add the content to be calculated to the shared content list when the preference matching value is greater than the preference matching threshold. The shared content list is stored in the database of the cloud server and is a list of content for team members to share or access. Team members can perform online operations and study on the shared content list.

12. A computing device, characterized in that: include: processor; A memory storing a computer program, which, when executed by a processor, implements the method for generating a shared content list for a team according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method for generating a shared content list for a team according to any one of claims 1 to 10 is implemented.

Images