Recommendation method and system for concerned users

Abstract

The embodiment of the invention provides an attention user recommendation method and system, and the method comprises the steps: obtaining first-degree attention users which are directly concerned by a to-be-recommended target user, and determining a preferred first-degree attention user from all the first-degree attention users based on the intimacy between the target user and each first-degree attention user; using the user directly concerned by each preferred first-degree concerned user as a second-degree concerned user of the target user, and screening alternative second-degree concerned users from all the second-degree concerned users based on a preset screening condition; for any alternative two-degree attention user, determining a comprehensive score of the alternative two-degree attention user; and according to the comprehensive score of each alternative two-degree attention user, determining a preferable two-degree attention user and recommending the preferable two-degree attention user to the target user. And recommending the preferable second-degree attention user to the target user, and providing richer, more diversified, more personalized and social transparency content recommendation for the target user, so that the user experience is enhanced, and the limitation of the prior art is overcome.

Description

Technical Field

[0001] The present invention relates to the field of content distribution, and in particular to a method and system for recommending users. Background Art

[0002] Personalized distribution of information flow has become one of the core functions of today's social networking platforms. Especially on platforms such as Weibo, the order of the follow flow is crucial to the user experience. Traditionally, the follow flow is mainly based on the accounts that the user directly follows - that is, the first-level follow network. This method works well when the user follows a large number of people, because there is a rich selection of content, making the information flow diverse and personalized.

[0003] However, when users follow fewer accounts, a major challenge is the limited content that can be distributed. The user experience may be affected due to the limited content that can be pushed. In addition, new or infrequent users usually need time to gradually discover and follow more bloggers, which is often a slow and time-consuming process.

[0004] When only relying on the user's direct social connections (i.e., first-degree relationships, i.e., accounts directly followed) or the user's personal behavior data, there is a lack of in-depth mining of the user's extended social network when providing recommendations. As a result, when the number of people followed by the user is small, the diversity and personalization of the recommended content is limited.

[0005] Therefore, it is particularly important to develop a method that can enrich the following stream without relying on users to actively expand their following lists. Summary of the invention

[0006] The embodiments of the present invention provide a method and system for recommending followed users, which can solve the problem in the prior art that only accounts directly followed by users can be relied upon to recommend content to users, resulting in limited diversity and personalization of recommended content.

[0007] To achieve the above-mentioned purpose, in a first aspect, an embodiment of the present invention provides a method for recommending users to be followed, comprising:

[0008] Obtaining first-degree attention users that the target user to be recommended directly follows, and determining preferred first-degree attention users from all first-degree attention users based on the intimacy between the target user and each first-degree attention user; the intimacy represents the interaction intensity and social closeness between the target user and the first-degree attention users;

[0009] Taking the users directly followed by each preferred first-degree following user as the second-degree following users of the target user, and screening the candidate second-degree following users from all the second-degree following users based on a preset screening condition;

[0010] For any candidate second-follow user, determine the comprehensive score of the candidate second-follow user according to the intimacy between the target user and the preferred first-follow user corresponding to the candidate second-follow user, the interest matching between the target user and the candidate second-follow user, and the influence of the candidate second-follow user;

[0011] According to the comprehensive score of each candidate second-follow user, a preferred second-follow user is determined and recommended to the target user.

[0012] In a second aspect, an embodiment of the present invention provides a recommendation system for following users, including:

[0013] A one-degree attention analysis unit is used to obtain one-degree attention users that the target user to be recommended directly follows, and determine a preferred one-degree attention user from all the one-degree attention users based on the intimacy between the target user and each one-degree attention user; the intimacy represents the interaction intensity and social closeness between the target user and the one-degree attention users;

[0014] A second-attention mining unit is used to take the users directly followed by each preferred first-attention user as second-attention users of the target user, and to filter candidate second-attention users from all second-attention users based on a preset filtering condition;

[0015] A comprehensive scoring unit, for determining a comprehensive score of any candidate second-followed user according to the intimacy between the target user and the preferred first-followed user corresponding to the candidate second-followed user, the interest matching between the target user and the candidate second-followed user, and the influence of the candidate second-followed user;

[0016] The recommendation unit is used to determine the preferred second-follow user according to the comprehensive score of each candidate second-follow user and recommend it to the target user.

[0017] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by a computer device, the computer device executes any one of the aforementioned methods for recommending users of interest.

[0018] The above technical solution has the following beneficial effects: expanding the user social network based on the second-degree attention relationship of the target user in the social network, that is, determining the preferred first-degree attention users of the target user according to the intimacy values ​​of the first-degree attention users directly followed by the target user; taking the users directly followed by each preferred first-degree attention user as the second-degree attention users of the target user, screening the alternative second-degree attention users from all the second-degree attention users based on preset screening conditions, and determining the preferred second-degree attention user according to the comprehensive score of any alternative second-degree attention user.

[0019] Preferred second-degree-followed users are mined from the second-degree-followed users of the target user, and the preferred second-degree-followed users are recommended to the target user, thereby more deeply mining the second-degree social network of the target user; the target user is thereby able to follow the content generated by the preferred second-degree-followed users, thereby achieving the goal of discovering new and potentially interesting users for the target user, thereby enriching the user's attention flow, and providing the target user with richer, more diverse, more personalized and socially transparent content recommendations, thereby enhancing the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 is a flow chart of a method for recommending users according to an embodiment of the present invention;

[0022] Figure 2 is a structural diagram of a user-focused recommendation system according to an embodiment of the present invention;

[0023] Figure 3 It is a sequential diagram of another method for recommending users according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] like Figure 1 As shown, in combination with an embodiment of the present invention, a method for recommending a user of interest includes:

[0026] S101: Obtaining first-degree following users that the target user to be recommended directly follows, and determining preferred first-degree following users from all first-degree following users based on the intimacy between the target user and each first-degree following user; the intimacy represents the interaction intensity and social closeness between the target user and the first-degree following users;

[0027] S102: taking the users directly followed by each preferred first-degree following user as second-degree following users of the target user, and screening candidate second-degree following users from all second-degree following users based on a preset screening condition;

[0028] S103: For any candidate second-followed user, determine a comprehensive score of the candidate second-followed user according to the intimacy between the target user and the preferred first-followed user corresponding to the candidate second-followed user, the interest matching between the target user and the candidate second-followed user, and the influence of the candidate second-followed user;

[0029] S104: Determine a preferred second-follow user according to the comprehensive score of each candidate second-follow user and recommend the user to the target user.

[0030] The user social network is expanded based on the secondary attention relationship of the target user in the social network, that is, the preferred first-degree attention users of the target user are determined according to the intimacy values ​​of the first-degree attention users directly followed by the target user; the users directly followed by each preferred first-degree attention user are taken as the second-degree attention users of the target user, and the alternative second-degree attention users are screened from all the second-degree attention users based on preset screening conditions, and the preferred second-degree attention user is determined according to the comprehensive score of any alternative second-degree attention user.

[0031] Preferred second-degree-followed users are mined from the second-degree-followed users of the target user, and the preferred second-degree-followed users are recommended to the target user, thereby more deeply mining the second-degree social network of the target user; the target user is thereby able to follow the content generated by the preferred second-degree-followed users, thereby achieving the goal of discovering new and potentially interesting users for the target user, thereby enriching the user's attention flow, and providing the target user with richer, more diverse, more personalized and socially transparent content recommendations, thereby enhancing the user experience, thereby overcoming the limitations of the prior art.

[0032] Preferably, S101: obtaining first-degree attention users that the target user to be recommended directly follows, and determining preferred first-degree attention users from all first-degree attention users based on the intimacy between the target user and each first-degree attention user, comprises:

[0033] S101-1: Acquire the first-degree follower that the target user directly follows through the follow list of the target user;

[0034] S101-2: sorting the first-time attention users in the attention list in descending order of the intimacy characteristic value according to the intimacy characteristic value between the target user and each first-time attention user;

[0035] S101-3: According to the total number of first-degree attention users that the target user directly follows, a set number of first-degree attention users whose intimacy characteristic values ​​are ranked first are determined as the preferred first-degree attention users; the set number is positively correlated with the total number.

[0036] Specifically: 1. Follow list analysis: collect the complete follow list of target user A, including all the first-time following users, such as all the first-time following bloggers.

[0037] 2. Intimacy sorting: Use the existing intimacy feature value for sorting. The intimacy feature value reflects the interaction intensity and closeness of social connection between target user A and each first-degree follower. The higher the value, the more the user prefers this first-degree follower. Sort the first-degree follower list of target user A by intimacy feature value to ensure that the first-degree follower with the closest relationship to target user A is given priority.

[0038] 3. Screening first-time following users (e.g., first-time following bloggers): Different retention criteria are applied to first-time following users according to the total number of followers of target user A, ensuring that only the most relevant first-time following users are retained in a large number of following lists, as shown in Table 1.

[0039] Table 1 Example of selecting preferred first-level users

[0040] Total number of followers The maximum number of users who have been following for a while 0-100 40 100-500 50 500-1000 60 More than 1000 70

[0041] Screening basis and purpose: The retention strategy for different numbers of followers is to ensure the efficiency of the recommendation system while avoiding information overload caused by too many followers.

[0042] The above operation can extract the most influential and relevant first-degree attention users from the user's huge attention network as an effective starting point for second-degree attention users, providing an accurate basis for subsequent second-degree attention relationship mining.

[0043] Preferably, S102: taking the users directly followed by each preferred first-degree following user as second-degree following users of the target user, and screening candidate second-degree following users from all second-degree following users based on a preset screening condition, including:

[0044] S102-1: Obtaining users who are directly followed by each preferred first-degree followed user through the follow list of each preferred first-degree followed user as second-degree followed users of the target user;

[0045] S102-2: Filter out the second-followed users that meet the preset screening conditions from all the second-followed users as candidate second-followed users; wherein the screening conditions include at least one of the following: a popularity condition, an influence condition, and an activity condition.

[0046] Specifically, 1. Mining of secondary following users: For each screened first-level following user (e.g., B1, B2, ..., B50), analyze their personal following list. Extract all second-level following bloggers (i.e., users followed by these first-level following users) from the following list of each first-level following user to form a second-level following user list containing C1, C2, ..., Cn.

[0047] 2. Initial screening and filtering of secondary followers: In order to ensure the quality and relevance of recommended content, secondary follow users are strictly screened and filtered, and screening conditions are set, such as: Popularity condition: It can be a big V certified account or a high-level account level to ensure the quality and credibility of recommended content; Influence condition: There should be a certain number of fans and a certain interaction rate with fans; Activity condition: It can be expressed by the number of information posted on the social platform.

[0048] Through screening, high-quality and highly relevant bloggers and second-follow users can be selected from a huge list of second-follow users. These second-follow users are most likely to attract target user A, thereby expanding the potential attention range of target user A and providing target user A with more diversified and personalized content recommendations.

[0049] Preferably, S103: for any candidate second-followed user, according to the intimacy between the target user and the preferred first-followed user corresponding to the candidate second-followed user, the interest matching between the target user and the candidate second-followed user, and the influence of the candidate second-followed user, determining the comprehensive score of the candidate second-followed user, including:

[0050] S103-1: For any candidate second-degree following user, calculating the intimacy score between the target user and the preferred first-degree following user corresponding to the candidate second-degree following user;

[0051] S103-2: Calculating the interest matching score between the target user and the candidate second-degree followed user;

[0052] S103-3: Calculating the influence score of the candidate second-followed user;

[0053] S103-4: Perform a weighted sum of the intimacy score, the interest matching score and the influence score to obtain a comprehensive score of the candidate second-degree follow-up user.

[0054] By calculating the comprehensive score, a multi-dimensional evaluation is formed, which can ensure that the recommended candidate second-degree attention C is not only highly matched with the personal interests of the target user A, but also has a significant influence in the social network. At the same time, the candidate second-degree attention C is followed by the first-degree attention user B who is liked by the target user A, thereby improving the overall quality of the recommended content and user satisfaction.

[0055] Preferably, S103-1: calculating the intimacy score between the target user and the preferred first-degree attention user corresponding to the candidate second-degree attention user, comprises:

[0056] Sum up the intimacy feature values ​​between each preferred first-degree attention user corresponding to the candidate second-degree attention user and the target user, and take the quotient of the sum of the intimacy feature values ​​and the number of preferred first-degree attention users corresponding to the candidate second-degree attention user as the intimacy score;

[0057] S103-2: Calculating the interest matching score between the target user and the candidate second-degree follow-up user, including:

[0058] Determine the tag weight of the field of interest of the target user, determine the ability weight of the candidate second-followed user to publish content in the field of interest of the target user, and use the product of the tag weight and the ability weight as the interest matching score between the target user and the candidate second-followed user;

[0059] S103-3: Calculating the influence score of the candidate second-followed user, including:

[0060] The influence score of the candidate second-followed user is obtained by taking a weighted sum of the number of fan users corresponding to the candidate second-followed user, the growth rate of the number of fan users, and the exposure of the content published by the candidate second-followed user.

[0061] Specifically, 1. The intimacy score between the target user and the user they once followed (ScoreAB):

[0062] From the target user A to the candidate second-degree attention user C, the target user A may go through the preferred first-degree attention users B1, B2, B3, B4, etc. The intimacy between the target user A and each preferred first-degree attention user B is calculated. The value range of ScoreAB is between 0-1, and the calculation formula is: ScoreAB = Σ(intimacy score) / number of preferred first-degree attention users; this formula takes into account the weighted average of the intimacy scores between the target user A and all preferred first-degree attention users B, among which the preferred first-degree attention users strengthen the influence of intimacy.

[0063] 2. Interest matching score between target user A and candidate second-follow users (ScoreAC):

[0064] Based on the matching of the interest tags of target user A and the content publishing ability tags of alternative second-follow user C, it reflects the degree of matching between the content publishing field of alternative second-follow user C and the interest field of alternative second-follow user A. The score reflects the potential interest matching degree of target user A to the content of alternative second-follow user C. ScoreAC is also between 0-1. The interest matching score is expressed as: ScoreAC = user interest tag weight * blogger content publishing ability tag weight.

[0065] For example, suppose target user A is interested in "technology" and "travel", and the interest tag weights are 0.6 and 0.7 respectively. Alternative secondary follower user C mainly publishes "technology" related content, and its content publishing ability tag weight is 0.8. Then the interest matching score is: ScoreAC = user interest tag weight * blogger content publishing ability tag weight = 0.6 * 0.8 = 0.48.

[0066] This calculation reflects the interest match between the target user A and the alternative second-follow user C in the field of "technology".

[0067] 3. Influence score of the candidate second-followed users (ScoreC):

[0068] The influence score of each candidate second-follow user C is calculated, and the calculation dimensions include: the number of fan users and the growth rate of the number of fan users.

[0069] ScoreC reflects the influence and popularity of the alternative second-follow user C on the social platform, and the value range is also between 0 and 1.

[0070] 4. Calculation of comprehensive scores of alternative users who have been followed twice:

[0071] Formula: Score total score = first coefficient * ScoreAB + second coefficient * ScoreAC + third coefficient * ScoreC, where first coefficient + second coefficient + ScoreC = 1.

[0072] For example, total score = 0.3*ScoreAB+0.3*ScoreAC+0.4*ScoreC.

[0073] Preferably, the method for recommending users of interest further includes:

[0074] S105: After the preferred second-followed user is recommended to the target user, the attention weight of the target user to the preferred second-followed user is adjusted according to the target user's feedback data on the content published by the preferred second-followed user, so as to adjust the recommendation strategy corresponding to the preferred second-followed user based on the attention weight.

[0075] By regularly using feedback data, we can better adapt to the changing preferences of target users, improve the accuracy of recommendations for preferred second-time users and improve the satisfaction of target users.

[0076] Preferably, in S105, adjusting the attention weight of the target user to the preferred second-followed user according to the feedback data of the target user on the content published by the preferred second-followed user, so as to adjust the recommendation strategy corresponding to the preferred second-followed user based on the attention weight, includes:

[0077] S105-1: Regularly collect the target user's attention information on the content published by the preferred second-followed user to form feedback data, wherein the feedback data includes: indicating no interest, no interaction (including no click), negative feedback, interaction and the number of interactions does not exceed the interaction number threshold, or interaction and the number of interactions exceeds the interaction number threshold; regularly extract the target user's feedback data on the preferred second-followed user from the offline table, and the feedback types include direct feedback (such as the user clicking the "not interested" button) and implicit feedback (for example, a preferred second-followed user C is exposed to the target user multiple times, but the target user does not interact with or click on its content).

[0078] S105-2: When the feedback data indicates no interest or no interaction, the attention weight of the target user to the preferred second-followed user is adjusted to zero, and the preferred second-followed user is not recommended to the target user again;

[0079] S105-3: when the feedback data is negative feedback, reducing the attention weight of the target user to the preferred second-followed user, and reducing the frequency of recommending the preferred second-followed user to the target user again;

[0080] S105-4: When the feedback data indicates that there is interaction and the number of interactions does not exceed the interaction number threshold, the attention weight of the target user to the preferred second-followed user is maintained, and the frequency of recommending the preferred second-followed user to the target user is maintained;

[0081] S105-5: When the feedback data is interactive and the number of interactions exceeds a preset interaction number threshold, increase the target user's attention weight for the preferred second-followed user, and increase the frequency of recommending the preferred second-followed user to the target user again.

[0082] Using the feedback data of target users to adjust and optimize recommendations, deleting some preferred second-follow users who are not of interest to the target users or reducing the recommendation frequency can better adapt to the changing preferences of target users, improve the accuracy of recommendations for preferred second-follow users and the satisfaction of target users.

[0083] Preferably, S104: determining a preferred second-follow user according to the comprehensive scores of the candidate second-follow users and recommending the user to the target user includes:

[0084] S104-1: sorting the candidate second-follow users in descending order according to the comprehensive score of each candidate second-follow user, generating a final candidate second-follow user list, and ensuring that the candidate second-follow user with the highest score, that is, the candidate second-follow user who is most likely to attract the target user A, is given priority;

[0085] S104-2: selecting a set number of candidate second-follow users ranked first in the comprehensive score as the preferred second-follow users, for example, selecting 2000 candidate second-follow users with the highest scores from the sorted candidate second-follow users list, and these candidate second-follow users constitute a final recommendation list, which will be presented to the target user A;

[0086] S104-3: recommending the preferred second-followed user to the target user through a recommendation interface, and displaying the recommendation reason for the preferred second-followed user on the recommendation interface, wherein the recommendation reason includes the preferred first-followed user corresponding to the preferred second-followed user.

[0087] In summary, taking Weibo as an example, the target user is a blogger. The sequential steps of the method for recommending users to follow in the embodiment of the present invention can be found in Figure 3 , the order includes the following:

[0088] 1. Users once paid attention to network analysis;

[0089] 2. Second focus on mining and filtering;

[0090] 3. Comprehensive score;

[0091] 4. Recommend and select bloggers you follow twice;

[0092] 5. User negative feedback and algorithm iteration.

[0093] like Figure 2As shown, in combination with an embodiment of the present invention, a recommendation system focusing on users includes:

[0094] A first-degree attention analysis unit 21 is used to obtain first-degree attention users that the target user to be recommended directly follows, and determine a preferred first-degree attention user from all first-degree attention users based on the intimacy between the target user and each first-degree attention user; the intimacy represents the interaction intensity and social closeness between the target user and the first-degree attention users;

[0095] A second attention mining unit 22, configured to take the users directly followed by each preferred first attention user as second attention users of the target user, and to filter candidate second attention users from all second attention users based on a preset filtering condition;

[0096] A comprehensive scoring unit 23 is used to determine a comprehensive score of any candidate second-followed user according to the intimacy between the target user and the preferred first-followed user corresponding to the candidate second-followed user, the interest matching degree between the target user and the candidate second-followed user, and the influence of the candidate second-followed user.

[0097] The recommendation unit 24 is used to determine the preferred second-follow user according to the comprehensive score of each candidate second-follow user and recommend it to the target user.

[0098] The user social network is expanded based on the secondary attention relationship of the target user in the social network, that is, the preferred first-degree attention users of the target user are determined according to the intimacy values ​​of the first-degree attention users directly followed by the target user; the users directly followed by each preferred first-degree attention user are taken as the second-degree attention users of the target user, and the alternative second-degree attention users are screened from all the second-degree attention users based on preset screening conditions, and the preferred second-degree attention user is determined according to the comprehensive score of any alternative second-degree attention user.

[0099] Preferred second-degree-followed users are mined from the second-degree-followed users of the target user, and the preferred second-degree-followed users are recommended to the target user, thereby more deeply mining the second-degree social network of the target user; the target user is thereby able to follow the content generated by the preferred second-degree-followed users, thereby achieving the goal of discovering new and potentially interesting users for the target user, thereby enriching the user's attention flow, and providing the target user with richer, more diverse, more personalized and socially transparent content recommendations, thereby enhancing the user experience, thereby overcoming the limitations of the prior art.

[0100] Preferably, the first attention analysis unit 21 is specifically used for:

[0101] Acquire the first-degree follower that the target user directly follows through the follow list of the target user;

[0102] According to the intimacy characteristic value between the target user and each of the first-followed users, the first-followed users in the follow list are sorted in descending order according to the intimacy characteristic value;

[0103] According to the total number of first-degree attention users that the target user directly follows, a set number of first-degree attention users whose intimacy characteristic values ​​are ranked first are determined as the preferred first-degree attention users; the set number is positively correlated with the total number.

[0104] Preferably, the secondary attention mining unit 22 is specifically used for:

[0105] Through the follow list of each preferred first-degree follower, users that each preferred first-degree follower directly follows are obtained as second-degree follower users of the target user;

[0106] Second-follow users that meet preset screening conditions are screened out from all second-follow users as candidate second-follow users; wherein the screening conditions include at least one of the following: a popularity condition, an influence condition, and an activity condition.

[0107] Preferably, the comprehensive scoring unit 23 includes:

[0108] An intimacy scoring subunit, for calculating, for any candidate second-degree attention user, an intimacy score between the target user and a preferred first-degree attention user corresponding to the candidate second-degree attention user;

[0109] An interest matching score subunit, used to calculate the interest matching score between the target user and the candidate second-degree follow-up user;

[0110] An influence scoring subunit, used to calculate the influence score of the candidate second-followed user;

[0111] The comprehensive scoring subunit is used to perform weighted summation on the intimacy score, the interest matching score and the influence score to obtain a comprehensive score of the candidate second-degree follow-up user.

[0112] Preferably, the intimacy scoring subunit is specifically used for:

[0113] Sum up the intimacy feature values ​​between each preferred first-degree attention user corresponding to the candidate second-degree attention user and the target user, and take the quotient of the sum of the intimacy feature values ​​and the number of preferred first-degree attention users corresponding to the candidate second-degree attention user as the intimacy score;

[0114] The interest matching scoring subunit is specifically used for:

[0115] Determine the tag weight of the field of interest of the target user, determine the ability weight of the candidate second-followed user to publish content in the field of interest of the target user, and use the product of the tag weight and the ability weight as the interest matching score between the target user and the candidate second-followed user;

[0116] The influence scoring subunit is used to:

[0117] The influence score of the candidate second-followed user is obtained by taking a weighted sum of the number of fan users corresponding to the candidate second-followed user, the growth rate of the number of fan users, and the exposure of the content published by the candidate second-followed user.

[0118] Preferably, the recommendation system for following users further includes:

[0119] An adjustment unit is used to adjust the target user's attention weight for the preferred second-followed user according to the target user's feedback data on the content published by the preferred second-followed user after recommending the preferred second-followed user to the target user, so as to adjust the recommendation strategy corresponding to the preferred second-followed user based on the attention weight.

[0120] Preferably, the adjustment unit is specifically used for:

[0121] Regularly collect the target user's attention information on the content published by the preferred second-follow user to form feedback data, the feedback data including: no interest, no interaction, negative feedback, interaction and the number of interactions does not exceed the interaction number threshold, or interaction and the number of interactions exceeds the interaction number threshold;

[0122] When the feedback data indicates that the user has no interest or has not interacted, the attention weight of the target user to the preferred second-followed user is adjusted to zero, and the preferred second-followed user is not recommended to the target user again;

[0123] When the feedback data is negative feedback, the attention weight of the target user to the preferred second-followed user is reduced, and the frequency of recommending the preferred second-followed user to the target user is reduced;

[0124] When the feedback data indicates that there is interaction and the number of interactions does not exceed the interaction number threshold, the attention weight of the target user to the preferred second-followed user is maintained, and the frequency of recommending the preferred second-followed user to the target user is maintained;

[0125] When the feedback data is interactive and the number of interactions exceeds a preset interaction number threshold, the attention weight of the target user to the preferred second-followed user is increased, and the frequency of recommending the preferred second-followed user to the target user again is increased.

[0126] Preferably, the recommendation unit 24 includes:

[0127] According to the comprehensive score of each candidate second-follow user, the candidate second-follow users are sorted in descending order according to the comprehensive score;

[0128] A set number of candidate second-follow users ranked first in the comprehensive score are used as the preferred second-follow users;

[0129] The preferred second-followed user is recommended to the target user through a recommendation interface, and a recommendation reason for the preferred second-followed user is displayed on the recommendation interface, wherein the recommendation reason includes the preferred first-followed user corresponding to the preferred second-followed user.

[0130] In conjunction with an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a computer device, the computer device executes any one of the aforementioned content recommendation methods.

[0131] The beneficial technical effects achieved by the embodiments of the present invention are as follows:

[0132] 1. Dig deeper into secondary social networks

[0133] Different from the existing technology that relies on the user's first-degree social connections or personal behavior data, the embodiments of the present invention more deeply mine the target user's second-degree social network and the accounts followed by the people followed by the target user, so as to discover new and potentially interesting preferred second-degree follow users for the target user, thereby enriching the target user's follow stream, which not only expands the scope of recommended content, but also improves the diversity and personalization of content, especially for those target users with fewer followers.

[0134] 2. Enhance transparency and personalization of recommendations

[0135] Clearly displaying why a preferred second-follow blogger C is recommended on the recommendation interface, such as "the bloggers B1 and B2 you follow also follow blogger C", enhances the personalization and social transparency of the recommendation. This clear display method makes the recommendation more persuasive and easy for target users to accept. This method not only shows the algorithm logic of the recommendation, but also makes the recommendation more humane, persuasive and personalized by showing the connections in the social network.

[0136] 3. Provide content recommendations based on social network structure

[0137] Compared with the recommendation systems of the prior art such as Facebook, Twitter, LinkedIn and YouTube (which mainly focus on first-degree social connections and user personal behaviors), the embodiments of the present invention focus more on providing content recommendations based on the social network structure of users, rather than being limited to first-degree follow relationships or personal behaviors. This is particularly effective for social media platforms such as Weibo, because it can deeply mine potential bloggers of interest based on the social network of the target user, rather than relying solely on the interests or behaviors that the target user has already demonstrated.

[0138] 4. Combining social network analysis with data-driven recommendations

[0139] It combines social network analysis and data-driven recommendation algorithms, dynamically adjusts re-recommendations through comprehensive scores and target user feedback data, and can dynamically adjust recommendations over time to adapt to changes in user preferences, significantly improving the exposure of social platforms such as Weibo relationship flows and the reading time of target users.

[0140] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process can be rearranged without departing from the scope of protection of the present disclosure. The attached method claims present the elements of the various steps in an exemplary order and are not intended to be limited to the specific order or hierarchy described.

[0141] In the above detailed description, various features are grouped together in a single embodiment to simplify the disclosure. This method of disclosure should not be interpreted as reflecting an intention that the embodiments of the claimed subject matter require more features than are clearly stated in each claim. On the contrary, as reflected in the appended claims, the invention is in a state of having less than all the features of the disclosed individual embodiments. Therefore, the appended claims are hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate preferred embodiment of the invention.

[0142] The disclosed embodiments are described above to enable any person skilled in the art to implement or use the present invention. Various modifications of these embodiments are obvious to those skilled in the art, and the general principles defined herein may also be applied to other embodiments without departing from the spirit and scope of the present disclosure. Therefore, the present disclosure is not limited to the embodiments given herein, but is consistent with the broadest scope of the principles and novel features disclosed in this application.

[0143] The above description includes examples of one or more embodiments. Of course, it is impossible to describe all possible combinations of components or methods for the purpose of describing the above embodiments, but it should be recognized by those skilled in the art that the various embodiments may be further combined and arranged. Therefore, the embodiments described herein are intended to cover all such changes, modifications and variations that fall within the scope of protection of the appended claims. In addition, with respect to the term "comprising" used in the specification or claims, the word is covered in a manner similar to the term "including", just as "including," is explained as a transitional word in the claims. In addition, any term "or" used in the specification of the claims is intended to mean "non-exclusive or".

[0144] Those skilled in the art may also understand that the various illustrative logical blocks, units, and steps listed in the embodiments of the present invention may be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly demonstrate the interchangeability of hardware and software, the various illustrative components, units, and steps described above have generally described their functions. Whether such functions are implemented by hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art may use various methods to implement the described functions for each specific application, but such implementation should not be understood as exceeding the scope of protection of the embodiments of the present invention.

[0145] The various illustrative logic blocks or units described in the embodiments of the present invention can be implemented or operated by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of the above. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any conventional processor, controller, microcontroller or state machine. The processor can also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a digital signal processor core, or any other similar configuration.

[0146] The steps of the method or algorithm described in the embodiments of the present invention can be directly embedded in hardware, a software module executed by a processor, or a combination of the two. The software module can be stored in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or other storage media of any form in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be arranged in an ASIC, and the ASIC can be arranged in a user terminal. Optionally, the processor and the storage medium can also be arranged in different components in the user terminal.

[0147] In one or more exemplary designs, the above functions described in the embodiments of the present invention can be implemented in hardware, software, firmware or any combination of the three. If implemented in software, these functions can be stored on a computer-readable medium, or transmitted in the form of one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media that facilitate the transfer of computer programs from one place to another. The storage medium can be any available medium that can be accessed by any general or special computer. For example, such computer-readable media can include but are not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store program codes in the form of instructions or data structures and other forms that can be read by general or special computers, or general or special processors. In addition, any connection can be appropriately defined as a computer-readable medium, for example, if the software is transmitted from a website site, server or other remote resource through a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wirelessly, such as infrared, wireless and microwave, it is also included in the defined computer-readable medium. The disk and disc include compact disk, laser disk, optical disk, DVD, floppy disk and blue-ray disk. Disks usually copy data magnetically, while discs usually copy data optically with lasers. The above combination can also be included in computer readable media.

[0148] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for recommending users, characterized in that: include: Obtaining first-degree attention users that the target user to be recommended directly follows, and determining preferred first-degree attention users from all first-degree attention users based on the intimacy between the target user and each first-degree attention user; the intimacy represents the interaction intensity and social closeness between the target user and the first-degree attention users; Taking the users directly followed by each preferred first-degree following user as the second-degree following users of the target user, and screening the candidate second-degree following users from all the second-degree following users based on a preset screening condition; For any candidate second-follow user, determine the comprehensive score of the candidate second-follow user according to the intimacy between the target user and the preferred first-follow user corresponding to the candidate second-follow user, the interest matching between the target user and the candidate second-follow user, and the influence of the candidate second-follow user; According to the comprehensive score of each candidate second-follow user, a preferred second-follow user is determined and recommended to the target user.

2. The method for recommending users of interest according to claim 1, characterized in that: The step of obtaining first-degree attention users that the target user to be recommended directly follows, and determining a preferred first-degree attention user from all the first-degree attention users based on the intimacy between the target user and each first-degree attention user, comprises: Acquire the first-degree follower that the target user directly follows through the follow list of the target user; According to the intimacy characteristic value between the target user and each of the first-followed users, the first-followed users in the follow list are sorted in descending order according to the intimacy characteristic value; According to the total number of first-degree attention users that the target user directly follows, a set number of first-degree attention users whose intimacy characteristic values ​​are ranked first are determined as the preferred first-degree attention users; the set number is positively correlated with the total number.

3. The method for recommending users of interest according to claim 1, characterized in that: The step of taking the users directly followed by each preferred first-degree following user as the second-degree following users of the target user, and screening candidate second-degree following users from all the second-degree following users based on a preset screening condition, comprises: Through the follow list of each preferred first-degree follower, users that each preferred first-degree follower directly follows are obtained as second-degree follower users of the target user; Second-follow users that meet preset screening conditions are screened out from all second-follow users as candidate second-follow users; wherein the screening conditions include at least one of the following: a popularity condition, an influence condition, and an activity condition.

4. The method for recommending users of interest according to claim 1, characterized in that: The step of determining, for any candidate second-followed user, a comprehensive score of the candidate second-followed user according to the intimacy between the target user and the preferred first-followed user corresponding to the candidate second-followed user, the interest matching degree between the target user and the candidate second-followed user, and the influence of the candidate second-followed user, comprises: For any candidate second-degree-followed user, calculating the intimacy score between the target user and the preferred first-degree-followed user corresponding to the candidate second-degree-followed user; Calculating the interest matching score between the target user and the candidate second-degree followed user; Calculate the influence score of the candidate second-followed user; The intimacy score, the interest matching score and the influence score are weightedly summed to obtain a comprehensive score of the candidate second-degree follow-up user.

5. The method for recommending users of interest according to claim 4, characterized in that: The calculating of the intimacy score between the target user and the preferred first-degree attention user corresponding to the candidate second-degree attention user comprises: Sum up the intimacy feature values ​​between each preferred first-degree attention user corresponding to the candidate second-degree attention user and the target user, and take the quotient of the sum of the intimacy feature values ​​and the number of preferred first-degree attention users corresponding to the candidate second-degree attention user as the intimacy score; The calculating the interest matching score between the target user and the candidate second-degree-followed user includes: Determine the tag weight of the field of interest of the target user, determine the ability weight of the candidate second-followed user to publish content in the field of interest of the target user, and use the product of the tag weight and the ability weight as the interest matching score between the target user and the candidate second-followed user; The calculating the influence score of the candidate second-followed user includes: The influence score of the candidate second-followed user is obtained by taking a weighted sum of the number of fan users corresponding to the candidate second-followed user, the growth rate of the number of fan users, and the exposure of the content published by the candidate second-followed user.

6. The method for recommending users of interest according to claim 1, characterized in that: Also includes: After the preferred second-follow user is recommended to the target user, the target user's attention weight for the preferred second-follow user is adjusted according to the target user's feedback data on the content published by the preferred second-follow user, so as to adjust the recommendation strategy corresponding to the preferred second-follow user based on the attention weight.

7. The method for recommending users of interest according to claim 6, characterized in that: The step of adjusting the target user's attention weight for the preferred second-followed user according to the target user's feedback data on the content published by the preferred second-followed user, so as to adjust the recommendation strategy corresponding to the preferred second-followed user based on the attention weight, includes: Regularly collect the target user's attention information on the content published by the preferred second-follow user to form feedback data, the feedback data including: no interest, no interaction, negative feedback, interaction and the number of interactions does not exceed the interaction number threshold, or interaction and the number of interactions exceeds the interaction number threshold; When the feedback data indicates that the user has no interest or has not interacted, the attention weight of the target user to the preferred second-followed user is adjusted to zero, and the preferred second-followed user is not recommended to the target user again; When the feedback data is negative feedback, the attention weight of the target user to the preferred second-followed user is reduced, and the frequency of recommending the preferred second-followed user to the target user is reduced; When the feedback data indicates that there is interaction and the number of interactions does not exceed the interaction number threshold, the attention weight of the target user to the preferred second-followed user is maintained, and the frequency of recommending the preferred second-followed user to the target user is maintained; When the feedback data is interactive and the number of interactions exceeds a preset interaction number threshold, the attention weight of the target user to the preferred second-followed user is increased, and the frequency of recommending the preferred second-followed user to the target user again is increased.

8. The method for recommending users of interest according to claim 1, characterized in that: The determining, based on the comprehensive scores of the candidate second-follow users, a preferred second-follow user and recommending the user to the target user comprises: According to the comprehensive score of each candidate second-follow user, the candidate second-follow users are sorted in descending order according to the comprehensive score; A set number of candidate second-follow users ranked first in the comprehensive score are used as the preferred second-follow users; The preferred second-followed user is recommended to the target user through a recommendation interface, and a recommendation reason for the preferred second-followed user is displayed on the recommendation interface, wherein the recommendation reason includes the preferred first-followed user corresponding to the preferred second-followed user.

9. A user-focused recommendation system, characterized in that: include: A one-degree attention analysis unit is used to obtain one-degree attention users that the target user to be recommended directly follows, and determine a preferred one-degree attention user from all the one-degree attention users based on the intimacy between the target user and each one-degree attention user; the intimacy represents the interaction intensity and social closeness between the target user and the one-degree attention users; A second-attention mining unit is used to take the users directly followed by each preferred first-attention user as second-attention users of the target user, and to filter candidate second-attention users from all second-attention users based on a preset filtering condition; A comprehensive scoring unit, for determining a comprehensive score of any candidate second-followed user according to the intimacy between the target user and the preferred first-followed user corresponding to the candidate second-followed user, the interest matching between the target user and the candidate second-followed user, and the influence of the candidate second-followed user; The recommendation unit is used to determine the preferred second-follow user according to the comprehensive score of each candidate second-follow user and recommend it to the target user.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a computer device, the computer device executes the method for recommending users of interest according to any one of claims 1 to 8.

Images