Resource pushing method and device, equipment and medium

Abstract

The invention provides a resource pushing method and device, equipment and a medium, and relates to the field of data processing, in particular to the technical field of big data, intelligent recommendation and live broadcast. According to the specific implementation scheme, first feature information of interaction users corresponding to all resource types is obtained; determining second feature information of each resource type according to the first feature information; determining third feature information of resource providing users corresponding to the resource types according to first interaction operations executed by the interaction users for the resources corresponding to the resource types; determining a to-be-recommended user corresponding to a target user according to fourth feature information, the second feature information and the third feature information of the target user; and pushing the resources created by the to-be-recommended user to the target user. According to the embodiment of the invention, the diversity of pushed resources can be increased, and the limitation of resource pushing is reduced.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing, specifically to the fields of big data, intelligent recommendation and live broadcast technology, and in particular to a resource push method, device, equipment and medium. Background Art

[0002] With the development of Internet technology, users can easily browse diverse network resources.

[0003] The recommendation system scores and ranks candidate resources (such as videos, pictures, texts, and live broadcasts) based on the user's historical behavior, interest preferences, and contextual information, thereby deciding which content will be displayed to the user first. Summary of the invention

[0004] The present disclosure provides a resource push method, device, equipment and medium.

[0005] According to one aspect of the present disclosure, a resource push method is provided, comprising:

[0006] Acquire first characteristic information of interactive users corresponding to each resource type;

[0007] Determining second characteristic information of each resource type according to the first characteristic information;

[0008] Determining third characteristic information of resource providing users corresponding to each resource type according to a first interactive operation performed by each interactive user on a resource corresponding to the resource type;

[0009] Determine, according to the fourth characteristic information of the target user, the second characteristic information and the third characteristic information, a user to be recommended corresponding to the target user, wherein there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user;

[0010] The resources created by the user to be recommended are pushed to the target user.

[0011] According to one aspect of the present disclosure, there is provided a resource pushing device, comprising:

[0012] A first feature determination module, used to obtain first feature information of interactive users corresponding to each resource type;

[0013] A second feature determination module, used to determine second feature information of each resource type according to the first feature information;

[0014] A third feature determination module, configured to determine third feature information of resource providing users corresponding to each resource type according to a first interaction operation performed by each interactive user on a resource corresponding to the resource type;

[0015] A recommended user determination module, used to determine a user to be recommended corresponding to the target user according to the fourth characteristic information of the target user, the second characteristic information and the third characteristic information, wherein there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user;

[0016] The resource push module is used to push the resources created by the user to be recommended to the target user.

[0017] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0018] at least one processor; and

[0019] a memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the resource pushing method described in any embodiment of the present disclosure.

[0021] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the resource pushing method described in any embodiment of the present disclosure.

[0022] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the resource pushing method described in any embodiment of the present disclosure is implemented.

[0023] The disclosed embodiments can increase the diversity of pushed resources and reduce the limitations of resource push.

[0024] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0026] Figure 1 is a flow chart of a resource push method disclosed in an embodiment of the present disclosure;

[0027] Figure 2 is a flow chart of another resource push method disclosed in an embodiment of the present disclosure;

[0028] Figure 3It is a scene diagram of a resource push method disclosed in an embodiment of the present disclosure;

[0029] Figure 4 It is a scene diagram of another resource push method disclosed in an embodiment of the present disclosure;

[0030] Figure 5 is a structural diagram of a resource push device disclosed in an embodiment of the present disclosure;

[0031] Figure 6 It is a block diagram of an electronic device according to the resource pushing method disclosed in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0032] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0033] Figure 1 This is a flow chart of a resource push method disclosed in an embodiment of the present disclosure. This embodiment can be applied to the case where resources that have not been browsed are pushed to users. The method of this embodiment can be executed by a resource push device, which can be implemented in software and / or hardware and is specifically configured in an electronic device with certain data computing capabilities, which can be a server device.

[0034] S101: Acquire first characteristic information of interactive users corresponding to each resource type.

[0035] The resource type is the type of resource, and the resource providing user can create at least one resource of the resource type. The interactive user can refer to a user who has interactive operations with the created resource. The interactive user corresponding to the resource type can refer to a user who has interactive operations with the resource created by the resource providing user belonging to the resource type. The interactive operations between the user and the resource can include the user browsing the resource, the user clicking on the resource, the user collecting the resource, and the user forwarding the resource.

[0036] In some embodiments, a resource type may correspond to no interactive user, and the resource type may be a generated type. In some embodiments, a resource type may correspond to at least one interactive user.

[0037] In some embodiments, the resource may refer to a live broadcast room, and the interactive user may refer to an audience. For another example, the resource may be a video, and the interactive user may refer to a user who browses the video. For another example, the resource may be a picture or text, and the interactive user may refer to a user who has browsed the picture or text.

[0038] The disclosed embodiment divides resources into resource types and pushes them according to the resource types, which can avoid the inability to accurately capture the user's interest preferences due to point-to-point push.

[0039] The first feature information is used to characterize the user's interest preference. The acquisition and use of the user's feature information involved in the embodiments of the present disclosure are known to the user and the user's consent is obtained.

[0040] S102: Determine second characteristic information of each resource type according to the first characteristic information.

[0041] The second feature information is used to characterize the user's interest preference for the resource type. For each resource type, the first feature information of the interactive user corresponding to the resource type can be fused to obtain the second feature information of the resource type.

[0042] S103: Determine third characteristic information of resource providing users corresponding to each resource type according to the first interactive operation performed by each interactive user on the resource corresponding to the resource type.

[0043] The existence of an interactive operation between an interactive user corresponding to a resource type and a resource of the resource type is equivalent to the existence of an interactive operation between the interactive user and the resource providing user who created the resource. In some embodiments, for each resource providing user, each interactive user corresponding to any resource type with which the resource providing user has an interactive operation is obtained, and based on the first feature information of these interactive users, the third feature information of these resource providing users is determined. At this time, the interactive users may not be limited to the interactive users corresponding to the resource type to which the resources of the resource providing user belong. In some embodiments, the interactive users with which the resource providing user has an interactive operation are determined among the interactive users corresponding to the resource type, and based on the first feature information of these interactive users, the third feature information of the resource providing users of the resource type is determined.

[0044] S104: Determine a user to be recommended corresponding to the target user according to the fourth characteristic information of the target user, the second characteristic information, and the third characteristic information, wherein there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user.

[0045] The target user is the user who needs to push resources. The user to be recommended may refer to a resource providing user who is potentially interested in the target user, and there is no valid interaction operation between the resource type corresponding to the user to be recommended and the target user. The user to be recommended is screened among the resource providing users. A valid interaction operation may be an operation in which the target user obtains data of any resource of the resource type. In some embodiments, a valid interaction operation may be an interaction behavior that lasts for a preset time, an interaction behavior that lasts for a preset frequency, or any interaction behavior. The valid interaction operation can be set according to demand.

[0046] The fourth characteristic information of the target user is used to indicate the user's personalized interest preference, the second characteristic information is used to indicate the user's interest preference for resource types, and the third characteristic information is used to indicate the user's interest preference for resource providers. Based on the second characteristic information, the third characteristic information, and the fourth characteristic information of the target user, the live broadcast types that the target user is interested in can be screened out, and the resource providers that the target user is interested in under the live broadcast types that the target user is interested in can be further screened out as users to be recommended.

[0047] S105: Push the resources of the to-be-recommended user corresponding to the target user to the target user.

[0048] The resources provided by the user to be recommended are determined as resources that the target user is interested in, and the resources provided by the user to be recommended are pushed to the target user.

[0049] According to the technical solution disclosed in the present invention, by obtaining the interactive users corresponding to the resource types, and determining the second characteristic information of the corresponding resource types according to the first characteristic information of the interactive users, and determining the third characteristic information of the resource providing user according to the interactive operations between the interactive users and the resource providing users, the second characteristic information, the third characteristic information and the fourth characteristic information of the target user are comprehensively considered to determine the user to be recommended, and the resource type to which the user to be recommended belongs does not have an effective interactive operation with the target user, so as to push new types of resources that have not been interacted with to the user, enrich the user's interaction types, break through the existing limitation of only being able to push resources of the resource types that have been displayed, increase the diversity of the pushed resource types, and push based on the resource type, it can reduce the poor push accuracy caused by the point-to-point push method of resources, and increase the push accuracy of resources of resource types of potential interest.

[0050] In an optional embodiment, the resource type is a live broadcast type, and the resource providing user is a live broadcast user.

[0051] The interactive user is the audience. The target user is the user of the live broadcast room to be pushed. The resource is the live broadcast resource created by the live broadcast user, and the live broadcast resource may include the live broadcast room, and the live broadcast resource may also include the live broadcast data stream.

[0052] In fact, as a relatively new and highly interactive content form, live streaming is often displayed less in hybrid recommendation systems (systems that recommend multiple types of resources such as videos, graphics, and live streaming) than traditional video and graphics content. In the live streaming push scenario, there is a clear situation where live streaming resources of the types that have been displayed are continuously recommended to users, and new types of live streaming resources rarely have the opportunity to be displayed to users. This tendency to frequently recommend live streaming types that users have already watched and narrow the range of recommended resource types is equivalent to limiting the range of recommended content, causing users to have an "information cocoon" effect in other unknown areas.

[0053] In addition, existing push methods usually require the target users to have a certain history of live broadcast viewing behavior as the basis for recommendation. For new users or users with less viewing history, due to the lack of sufficient data support, it is impossible to accurately capture their interest preferences, resulting in inaccurate recommendation results.

[0054] The disclosed embodiments make live broadcast recommendations without relying on the user's historical live broadcast viewing behavior, and can recommend live broadcast types that the user has not watched, thereby effectively expanding the breadth of live broadcast types for the user, improving the accuracy of the recommendations while expanding the range of categories to which the recommended videos belong, and further improving the user experience.

[0055] In an optional embodiment, the resource push method further includes: acquiring a resource providing user who has interactive operations with the interactive user; and generating the first feature information according to identification information of the interactive user and identification information of each resource providing user with interactive operations.

[0056] For each interactive user, obtain the resource providing user with whom the interactive user has performed interactive operations, as well as obtain the identification information of the interactive user and the identification information of the resource providing user, establish a relationship between the two pieces of identification information, and generate the first feature information of the interactive user based on the identification information of the interactive user and the identification information of the resource providing user with which the relationship is established.

[0057] In some embodiments, a graph may be generated based on the identification information of the interactive users and the established relationships, wherein the nodes in the graph include the interactive users and the resource providing users, etc. The edges in the graph indicate the existence of interactive operations, and the edges between two nodes indicate the existence of interactive operations between the users represented by the two nodes. A graph model is used to determine the vector representation of each node based on the established graph, and the vector representation of the node as the interactive user is determined as the first feature information of the interactive user. Among them, the target user is a node in the graph, and accordingly, the fourth feature information of the target user can also be determined in the same way.

[0058] By obtaining the identification information of the interactive user, the identification information of the resource providing user and whether there is an interactive operation, the structure and relationship between the interactive user and the resource providing user can be captured, and based on this, the characteristic information describing the potential interests of the interactive user can be determined, so as to accurately determine the characteristic information of the group of people interested in the live broadcast type, and so on, the fourth characteristic information of the target user can be determined, thereby increasing the accuracy of resource push.

[0059] Figure 2 It is a flowchart of another resource push method disclosed in an embodiment of the present disclosure, which is further optimized and expanded based on the above technical solution and can be combined with the above optional implementation methods. According to the first characteristic information, the second characteristic information of each resource type is determined, which is specifically: according to the user vector of each interactive user of the resource type, each interactive user is clustered to obtain the type vector of the cluster center of the resource type, and the second characteristic information includes the type vector.

[0060] S201: Obtain first characteristic information of interactive users corresponding to each resource type.

[0061] S202: Cluster the interactive users according to the user vectors of the interactive users of the resource type to obtain a type vector of a cluster center of the resource type, wherein the second feature information includes the type vector.

[0062] Among them, the first characteristic information of the interactive user includes the user vector of the interactive user. The interactive users are divided according to different resource types. For the interactive users corresponding to a resource type, all the interactive users of the resource type are clustered, and the cluster center is determined therefrom, and the cluster center is used to represent the resource type. The user vectors of each interactive user of the resource type are fused to obtain the type vector of the cluster center, that is, the type vector of the resource type, and used as the second characteristic information of the resource type. In fact, the interactive users of the resource type all have interactive operations with the resources of the resource type, indicating that the characteristics of the resource type exist in these interactive users. These interactive users are aggregated, and the preference interest characteristics of the interactive users for the resource type are extracted, thereby obtaining the second characteristic information of the resource type.

[0063] S203: Determine third characteristic information of resource-providing users corresponding to each resource type according to the first interactive operations performed by each interactive user on the resource corresponding to the resource type.

[0064] S204: Determine a user to be recommended corresponding to the target user according to the fourth characteristic information of the target user, the second characteristic information, and the third characteristic information, wherein there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user.

[0065] S205: Push the resources created by the user to be recommended to the target user.

[0066] According to the technical solution disclosed in the present invention, by clustering the interactive users of resource types to obtain cluster centers and type vectors, and using the cluster centers to represent resource types, the interests, preferences and needs of users who interact with resource types can be accurately mined, and resource types can be positioned as common features of interest preference user groups, so as to match resource types with target users, implement push judgments without historical behavior, and improve push accuracy.

[0067] In an optional embodiment, the resource push method also includes: determining reliable users from each of the interactive users based on the user vector and the type vector of each of the interactive users; determining third characteristic information of resource providing users corresponding to each of the resource types based on the first interactive operation performed by each of the interactive users on the resource corresponding to the resource type, including: determining the interaction probability of each of the reliable users with each of the resource providing users based on the second interactive operation performed by the reliable user on the resource providing user of the resource type and the operation type of the second interactive operation; determining the third characteristic information based on the user vector of the reliable user and the interaction probability.

[0068] Among them, reliable users can refer to representative users who are interested in the live broadcast type among interactive users. Interactive users can be screened according to their confidence. According to the user vector of the valid users of the resource type and the type vector of the cluster center of the resource type, the distance between each valid user of the resource type and the cluster center is determined, and n valid users with close distances can be selected according to the distance to determine the reliable users, and the distance represents the similarity between the valid users and the cluster center. In some examples, the top 10 valid users with the closest distance can be screened out and determined as reliable users.

[0069] Among them, the interaction probability of a reliable user for a resource providing user may refer to the probability of an interaction operation between the reliable user and the resource provided by the resource providing user. Different operation types have different influences on the interaction probability, and different weights corresponding to different operation types may be preset. In some embodiments, the weight corresponding to the operation type and the number of interaction operations may be weighted and summed to calculate the interaction probability.

[0070] For each resource providing user, the user vectors of reliable users who have interactive operations with the resource providing user can be fused according to the interaction probability of each reliable user with respect to the resource providing user to obtain the user vector of the resource providing user as the third feature information of the resource providing user. In some embodiments, the user vectors of reliable users are weightedly fused using the interaction probability as a weight to obtain the third feature information of the resource providing user.

[0071] In some examples, the operation type may include viewing, interaction, and resource interaction. Count the number of different a posteriori behaviors of reliable users on the live broadcast user, including: valid viewing number view_num, interaction number: hudong_num, and gift giving number: gift_num. Set different weights for different operation types, for example, the weight of viewing is w_a, the weight of interaction is w_b, and the weight of resource interaction is w_c. Perform weighted summation on the a posteriori behavior times to obtain the total number total_num:

[0072] total_num=

[0073] max(view_num,5)*w_a+max(hudong_num,5)*w_b+max(gift_num,5)*w_c

[0074] Among them, w_a+w_b+w_c=1.0, generally w_c>w_b>w_a.

[0075] Calculate the interaction probability star_score: star_score = sigmoid(total_num). Since sigmoid(5) = 0.99 is close to 1.0, the maximum number of posterior behaviors is 5.

[0076] Calculate the user vector of the resource provider, that is, the user vector of the live broadcast user: use the weighted average of the user vectors user_embedding of reliable users to obtain the user vector star_embedding of the resource provider:

[0077]

[0078] Where n is the number of reliable users.

[0079] By screening the interactive users of the resource type according to the cluster center of the resource type to obtain reliable users, and screening out more representative reliable users, and calculating the third feature information based on the reliable users, the effective information in the third feature information can be increased, so as to determine the users to be recommended based on the third feature information, which can improve the accuracy of the users to be recommended.

[0080] In an optional embodiment, determining the user to be recommended corresponding to the target user based on the fourth characteristic information, the second characteristic information and the third characteristic information of the target user includes: obtaining the first interest information of the target user for each of the resource types based on the fourth characteristic information and the third characteristic information; obtaining the target resource type with which the target user has a valid interactive operation; determining the target type corresponding to the target user based on the first interest information and the target resource type; obtaining the second interest information of the target user for a resource provider of the target type based on the fourth characteristic information and the third characteristic information; determining the recommendation probability based on at least the second interest information; and determining the user to be recommended based on the recommendation probability.

[0081] The first interest information is used to describe the target user's interest in the resource of the resource type. The existence of a valid interaction operation between the target user and the resource type may refer to the existence of a valid interaction operation between the target user and any resource of the resource type. A valid interaction operation may refer to an interaction operation in which the target user intends to perceive any resource of the resource type. An invalid interaction operation may refer to an interaction operation in which the target user has no intention to perceive all resources of the resource type.

[0082] In some embodiments, whether the user has the intention to perceive the resource type can be judged by the perception time period, perception duration and operation type of the interactive operation. In some examples, the perception time period is within 2 weeks before the current time, the perception duration is 30 seconds, and the operation type of the interactive operation can be browsing. Exemplarily, the existence of a valid interactive operation between the target user and the resource type can be that the target user has watched at least one live broadcast room of the resource type for 30 seconds within 2 weeks. For another example, the existence of an invalid interactive operation between the target user and the resource type can be that within a month, the target has watched the live broadcast room of the resource type for less than 30 seconds.

[0083] Among them, the interaction type may refer to the resource type to which the target user has historically intended to perceive the resource. The target type may refer to the resource type that the target user has not historically perceived. According to the first interest information of each resource type, the resource types that the target user is interested in are screened out, and the interaction types are eliminated from the resource types of interest, so that the target types that the user has not perceived and is interested in can be obtained. The second interest information is used to describe the target user's interest in the resource provider user, and these resource providers belong to the target type. The recommendation probability is used to describe the degree to which the resource provider user recommends to the target user.

[0084] According to the fourth characteristic information of the target user and the third characteristic information of the target type, the user to be recommended corresponding to the target user is screened from the resource providing users of the target type. Optionally, according to the fourth characteristic information of the target user and the third characteristic information of the target type, the second interest information of the target user for the resource providing users of the target type can be calculated.

[0085] Among them, the recommendation probability can be determined only based on the second interest information, or the second interest information can be combined with other information to jointly determine the recommendation probability. In some embodiments, the recommendation probability of the target user to the resource provider of the target type can be determined based on at least one of the second interest information of the target user for each resource provider of the target type, the first interest information of the target user for the target type, and the interaction probability of each resource provider of the target type. In some embodiments, the second interest information of the target user for the resource provider of the target type is determined as the recommendation probability of the target user to the resource provider of the target type; in some embodiments, the recommendation probability of the target user to the resource provider of the target type is determined based on the second interest information of the target user for each resource provider of the target type, the first interest information of the target user for the target type, and the interaction probability of each resource provider of the target type. The resource provider with the highest recommendation probability is determined as the user to be recommended, or the resource provider with the highest recommendation probability is determined as the user to be recommended, or the resource provider with a recommendation probability greater than or equal to a preset probability threshold is determined as the user to be recommended.

[0086] By calculating the first interest information of the target user for the resource type based on the fourth characteristic information of the target user and the second characteristic information of the resource type, and based on the interaction type in which the target user has effective interaction operations, filtering out the target types that meet the invalid interaction conditions from the resource types, and filtering out the recommended users corresponding to the target user from the resource providing users of the target type, the target types of interest can be preferentially filtered out, and then the recommended users can be filtered out from a small range of target types, thereby improving the screening accuracy of the recommended users.

[0087] In an optional embodiment, the recommendation probability includes: a fusion probability of the second interest information, the first interest information of the target user for the target type, and the interaction probability of each resource providing user of the target type.

[0088] Among them, the second interest information represents the target user's interest in the individual user. The first interest information represents the target user's interest in the resource type. The interaction probability represents the popularity of the resource provider to the user. The fusion probability may refer to the fusion result of the second interest information, the first interest information and the interaction probability. In some embodiments, the weighted sum of the second interest information, the first interest information and the interaction probability is calculated to obtain the fusion probability. In some embodiments, the second interest information, the first interest information and the interaction probability may be spliced ​​to obtain the fusion probability. The fusion probability may be used as a recommendation probability. Alternatively, the fusion probability may be combined with other information to form a recommendation probability. Other information may be set as needed.

[0089] By fusing the three probabilities to obtain the fusion probability and determine the recommendation probability, the recommendation probability can be calculated in multiple dimensions to increase the accuracy of the recommendation probability.

[0090] In an optional embodiment, determining the recommendation probability at least based on the second interest information includes: obtaining weight configuration information, and determining the type weight, popularity weight and personalized weight from the weight configuration information; obtaining a first probability based on the product of the type weight and the first interest information of the target user for the target type; obtaining a second probability based on the product of the popularity weight and the interaction probability; obtaining a third probability based on the product of the personalized weight and the second interest information; and obtaining the recommendation probability based on the sum of the first probability, the second probability and the third probability.

[0091] Among them, the type weight is used to determine the degree of influence of the target user's first interest information for the target type on the recommendation probability. The first probability may refer to the adjusted interest level of the target user for the target type. The popular probability is used to determine the degree of influence of the interaction probability of the resource providing user on the recommendation probability. The second probability may refer to the adjusted popularity of the resource providing user. The personalized weight is used to determine the degree of influence of the target user's second interest information for the resource providing user on the recommendation probability. The first probability may refer to the adjusted interest level of the target user for the resource providing user.

[0092] In some examples, the recommendation probability final_score is calculated as follows:

[0093] final_score=w_cate*cate_score+w_star*star_score+w_user*user_score

[0094] Among them, w_cate is the type weight, cate_score is the first interest information, w_star is the popularity weight, star_score is the interaction probability, w_user is the personalized weight, and user_score is the second interest information. w_cate+w_star+w_user=1.0, increasing w_cate means strengthening the role of resource types, increasing w_star means tending to recommend popular resources in each resource type, and increasing w_user pays more attention to the personalized interests of users. The weights can be set as needed.

[0095] By setting corresponding weights for the first interest information, the second interest information and the interaction probability respectively, the influence of each probability on the recommendation probability can be flexibly adjusted, so that the content of the recommended resources can be flexibly adjusted according to the recommendation requirements, thereby increasing the flexibility of the recommended resources.

[0096] In an optional embodiment, there is no interaction operation between the target user and each resource providing user of the target type.

[0097] Among them, the target user has no interactive operation with all resource providers of the target type, indicating that the target user has no interactive operation with the target type. The non-interaction of resource providers belonging to the target type is used as the basis for screening the target type, so as to accurately determine the resources to be pushed from the resource types that the target user has not interacted with, thereby reducing the limitations of the push scope.

[0098] In an optional embodiment, there is a valid interactive operation between the interactive user and the resource providing user of the resource type, wherein the valid interactive operation includes an interactive operation that lasts for a valid period within a valid time period.

[0099] The valid time period defines the time when the interactive operation occurs, and the valid duration defines the duration of the interactive operation. In some examples, the valid duration is 30 seconds, and the valid time period is one month before the current time.

[0100] In some embodiments, the effective time periods of different resource types are independent of each other, and the effective time periods of different resource types can be different or the same. The effective durations of different resource types are independent of each other, and the effective durations of different resource types can be different or the same. In addition, independent operation types can be set for different resource types. The number of interactive operations is at least one.

[0101] In some examples, the resource is a live broadcast resource, and the resource type is a live broadcast type. For example, the valid interactive operation of the game type is: a viewing operation lasting 1 minute within 2 weeks. For another example, the valid interactive operation of the popular science type is: a viewing operation lasting 10 minutes within half a year. For another example, the valid interactive operation of the beauty type is: a viewing operation lasting 5 seconds within a year, and a collection operation lasting 1 week within a year.

[0102] By limiting the effective time period and effective duration for effective interactive operations, the granularity of judging interactive operations can be increased, effective interactive operations can be accurately judged, and the interference of redundant operations can be reduced. Based on effective interactive operations, user groups that accurately represent resource types can be screened out, thereby improving the accuracy of recommended resources.

[0103] In an optional embodiment, after determining the third characteristic information of the resource providing user corresponding to each resource type according to the first interactive operation performed by each interactive user on the resource corresponding to the resource type, the method further includes: obtaining, from each resource providing user of the resource type, a user to be adjusted who does not belong to the resource type according to the second characteristic information and the third characteristic information; adjusting the resource type to which the user to be adjusted belongs; and adjusting each corresponding interactive user, each second characteristic information and the third characteristic information according to the adjusted resource type.

[0104] In fact, the correspondence between the resource providing user and the resource type is manually set, and there may be errors in the correspondence, which can be corrected according to the third characteristic information and the second characteristic information.

[0105] The user to be adjusted may refer to a resource providing user who is classified into an incorrect resource type. For a resource type, the similarity between the resource type and the resource providing user is calculated based on the second feature information of the resource type and the third feature information of the resource providing user of the resource type. The average similarity between the resource type and all resource providing users of the resource type is counted, and the ratio between the similarity of each resource providing user and the average similarity is calculated. According to the ratio of a resource providing user, it is determined whether the resource providing user is a user to be adjusted whose resource type is incorrectly classified.

[0106] In some embodiments, based on the type vector cate_embedding of the resource type and the user vector star_embedding of the resource providing user, the distance star_distance between the resource type and the resource providing user is calculated as the similarity between the resource type and the resource providing user. The average distance avg_distance between the resource type and all resource providing users of the resource type is counted as the average similarity. If the ratio of the distance star_distance of a resource providing user to the average distance avg_distance is greater than or equal to the ratio threshold thresh, that is, star_distance / avg_distance>thresh. It is determined that the resource type classification of the resource providing user is wrong. Wherein thresh is an empirical value, which can be obtained based on business data analysis. The larger the thresh, the greater the possibility that the selected resource type classification is wrong.

[0107] The similarity between the user to be adjusted and each other resource type can be determined based on the characteristic information of the user to be adjusted and the second characteristic information of other resource types, and the resource type with the highest similarity to the user to be adjusted can be screened out based on the similarity between the user to be adjusted and each other resource type, and determined as the resource type to which the user to be adjusted belongs after adjustment. In addition, the operation and maintenance user can be prompted to perform manual review and adjustment.

[0108] Since the resource providing user of the resource type is modified, it is necessary to obtain the resource type that needs to be modified, that is, the resource type of the added resource providing user and the resource type of the deleted resource providing user. For the resource providing user of the modified resource type, re-determine the interactive user, the second characteristic information of the modified resource type, and the third characteristic information of the resource providing user of the modified resource type. And based on the modified characteristic information, determine whether there is a user to be adjusted among the resource providing users of the modified resource type. If there is still a user to be adjusted, repeat the above adjustment process. If there is no user to be adjusted. Based on the modified characteristic information, interactive users and resource providing users, determine the user to be recommended for the target user.

[0109] By correcting the affiliation between the resource providing user and the resource type to which they belong, the accuracy of the affiliation between the resource providing user and the resource type can be improved, thereby screening out accurate users to be recommended.

[0110] In a scenario, such as Figure 3-Figure 4 A scene diagram of a resource push method disclosed in an embodiment of the present disclosure is shown.

[0111] like Figure 3-Figure 4 As shown, the resource push method applied to the live broadcast field includes:

[0112] S301, obtaining the live broadcast type.

[0113] Live streaming users are divided into different types according to the live streaming content. For example, live streaming types include: appearance type, game type... outdoor type.

[0114] S302: Mining interactive users.

[0115] Get the interactive users who have interactive operations with the live broadcast resources of each live broadcast type as seed users. Cluster the interactive users of the live broadcast type to obtain the cluster center. For example, the cluster center may include: the cluster center of the appearance type, the cluster center of the game type... the cluster center of the outdoor type. According to the user vector of the interactive user, obtain the type vector of the cluster center and use it as the type vector of the live broadcast type.

[0116] S303: Acquire reliable user interaction resources and provide them to the user.

[0117] According to the cluster center, reliable users are screened from the interactive users. Reliable users may include: reliable users of appearance type, reliable users of game type...reliable users of outdoor type. Based on the user vectors of reliable users corresponding to the live broadcast type and the interactive operations of reliable users on the live broadcast users, the user vectors of each live broadcast user are determined. The user vectors may include: user vectors of live broadcast users of appearance type, user vectors of live broadcast users of game type...user vectors of live broadcast users of outdoor type.

[0118] S304: Adjust the live broadcast type.

[0119] According to the type vector of the live broadcast type and the user vector of the live broadcast user, it is determined whether the live broadcast user is correctly classified into the current live broadcast type, and the live broadcast type to which the wrong live broadcast user belongs is adjusted. The live broadcast type is reclassified for the live broadcast user, the adjusted live broadcast type is obtained, and interactive users are re-mined based on the reclassified live broadcast type.

[0120] S305: Screen potential target types of interest.

[0121] The live broadcast types that meet the invalid interaction condition with the target user are screened out from the live broadcast types, and the target type is screened out from the live broadcast types that meet the invalid interaction condition according to the user vector of the target user and the type vector of the screened live broadcast types. The target type may include: potential interested target type 1, potential interested target type 2, ... potential interested target type n.

[0122] S306: Screening live broadcast resources of target types of potential interest.

[0123] According to the user vector of the target user and the type vector of the target type, the probability of the target user being interested in the target type is calculated; according to the user vector of the target user and the user vectors of the live users of the target type, the probability of the target user being interested in each live user of the target type is calculated; according to the interactive operations between the reliable users of the target type and the live users of the target type, the interaction probability of the live users of the target type is determined.

[0124] Calculate the probability of the target user being interested in the target type of live broadcast users, the probability of the target user being interested in the target type, and the probability of interaction with each live broadcast user of the target type, and calculate the recommendation probability of the target user for the target type of live broadcast users. For example, the recommendation probability may include: the recommendation probability of the appearance type live broadcast users, the recommendation probability of the game type live broadcast users...the recommendation probability of the outdoor type live broadcast users. Filter out the users to be recommended from the live broadcast users, and push the live broadcast resources of the users to be recommended to the target users.

[0125] The disclosed embodiments do not rely on the user's historical live broadcast viewing behavior to make live broadcast recommendations, and can recommend live broadcast categories that the user has not watched, thereby effectively expanding the range of live broadcast types that the user watches, increasing the breadth of recommended live broadcast types, and improving user experience.

[0126] According to an embodiment of the present disclosure, Figure 5 1 is a structural diagram of a resource push device in an embodiment of the present disclosure, and the embodiment of the present disclosure is applicable to the case where resources that have not been browsed are pushed to a user. The device is implemented by software and / or hardware, and is specifically configured in an electronic device with certain data computing capabilities.

[0127] like Figure 5 The resource push device 500 shown in the figure comprises: an interactive user acquisition module 501, a second feature determination module 502, a third feature determination module 503, a recommended user determination module 504 and a resource push module 505.

[0128] An interactive user acquisition module 501 is used to acquire first characteristic information of interactive users corresponding to each resource type;

[0129] A second feature determination module 502, configured to determine second feature information of each resource type according to the first feature information;

[0130] A third feature determination module 503, configured to determine third feature information of resource providing users corresponding to each resource type according to the first interaction operation performed by each interactive user on the resource corresponding to the resource type;

[0131] The recommended user determination module 504 is used to determine a user to be recommended corresponding to the target user according to the fourth characteristic information of the target user, the second characteristic information and the third characteristic information, and there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user;

[0132] The resource push module 505 is used to push the resources created by the user to be recommended to the target user.

[0133] According to the technical solution disclosed in the present invention, by obtaining the interactive users corresponding to the resource types, and determining the second characteristic information of the corresponding resource types according to the first characteristic information of the interactive users, and determining the third characteristic information of the resource providing user according to the interactive operations between the interactive users and the resource providing users, the second characteristic information, the third characteristic information and the fourth characteristic information of the target user are comprehensively considered to determine the user to be recommended, and the resource type to which the user to be recommended belongs does not have an effective interactive operation with the target user, so as to push new types of resources that have not been interacted with to the user, enrich the user's interaction types, break through the existing limitation of only being able to push resources of the resource types that have been displayed, increase the diversity of the pushed resource types, and push based on the resource type, it can reduce the poor push accuracy caused by the point-to-point push method of resources, and increase the push accuracy of resources of resource types of potential interest.

[0134] Optionally, the second feature determination module 502 includes:

[0135] The interactive clustering unit is used to cluster the interactive users according to the user vectors of the interactive users of the resource type to obtain the type vector of the cluster center of the resource type, and the second feature information includes the type vector.

[0136] Optionally, the resource pushing device further includes:

[0137] A reliable screening module, used for determining a reliable user from each of the interactive users according to the user vector and the type vector of each of the interactive users;

[0138] Wherein, the third feature determination module 503 includes:

[0139] An interaction probability calculation unit, configured to determine the interaction probability of each of the reliable users with respect to each of the resource providing users according to the second interaction operation performed by the reliable user with respect to the resource providing user of the resource type and the operation type of the second interaction operation;

[0140] The third feature determination unit is used to determine the third feature information according to the user vector of the reliable user and the interaction probability.

[0141] Optionally, the recommended user determination module 504 includes:

[0142] A first interest calculation unit, configured to obtain first interest information of the target user for each resource type according to the fourth feature information and the third feature information;

[0143] An interaction type determination unit, configured to obtain a target resource type for which the target user has a valid interaction operation;

[0144] a resource type screening unit, configured to determine a target type corresponding to the target user according to the first interest information and the target resource type;

[0145] A second interest calculation unit, configured to obtain second interest information of the target user providing the user with respect to the resource of the target type according to the fourth feature information and the third feature information;

[0146] a recommendation probability calculation unit, configured to determine a recommendation probability at least according to the second interest information;

[0147] The recommended user determination unit is used to determine the user to be recommended according to the recommendation probability.

[0148] Optionally, the recommendation probability includes: a fusion probability of the second interest information, the first interest information of the target user for the target type, and the interaction probability of each resource providing user of the target type.

[0149] Optionally, the recommendation probability calculation unit is used to:

[0150] Obtain weight configuration information, and determine type weight, hot weight and personalized weight from the weight configuration information;

[0151] Calculate the product of the type weight and the first interest information of the target user for the target type to obtain a first probability;

[0152] Calculate the product of the hot weight and the interaction probability to obtain a second probability;

[0153] Obtaining a third probability according to a product of the personalized weight and the second interest information;

[0154] The sum of the first probability, the second probability and the third probability is calculated to obtain the recommendation probability.

[0155] Optionally, there is no interaction operation between the target user and each resource providing user of the target type.

[0156] Optionally, there is a valid interaction operation between the interactive user and the resource providing user of the resource type, wherein the valid interaction operation includes an interaction operation that lasts for a valid period within a valid time period.

[0157] Optionally, the resource pushing device further includes:

[0158] an adjustment user screening module, configured to obtain, from among the resource providing users of the resource type, users to be adjusted that do not belong to the resource type, according to the second characteristic information and the third characteristic information, after determining the third characteristic information of the resource providing users corresponding to the resource type according to the first interaction operation performed by each of the interactive users on the resource corresponding to the resource type;

[0159] A resource type adjustment module, used to adjust the resource type to which the user to be adjusted belongs;

[0160] The characteristic information adjustment module is used to adjust each corresponding interactive user, each second characteristic information and the third characteristic information according to the adjusted resource type.

[0161] Optionally, the resource type is a live broadcast type, and the resource providing user is a live broadcast user.

[0162] Optionally, the resource pushing device further includes:

[0163] Provide a user acquisition module for acquiring resource providing users who have interactive operations with interactive users;

[0164] The feature information generating module is used to generate the first feature information according to the identification information of the interactive user and the identification information of each resource providing user with interactive operations.

[0165] The resource push device can execute the resource push method provided by any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects for executing the resource push method.

[0166] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0167] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0168] Figure 6A schematic area diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0169] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the device 600 instructions can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0170] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0171] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the resource push method. For example, in some embodiments, the resource push method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the resource push method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the resource push method in any other appropriate manner (e.g., by means of firmware).

[0172] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard objects (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0173] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, implements the functions / instructions specified in the flow chart and / or the area diagram. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0174] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0175] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0176] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0177] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services. The server may also be a server of a distributed system, or a server combined with a blockchain.

[0178] Artificial intelligence is a discipline that studies how to use computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.). It includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, machine learning / deep learning technology, big data processing technology, knowledge graph technology, and other major directions.

[0179] Cloud computing refers to a technology system that uses network access to elastically scalable shared physical or virtual resource pools. Resources can include servers, instruction systems, networks, software, applications, and storage devices, and can be deployed and managed on demand and in a self-service manner. Cloud computing technology can provide efficient and powerful data processing capabilities for the application of technologies such as artificial intelligence and blockchain, as well as model training.

[0180] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution provided by this disclosure can be achieved, and this document does not limit this.

[0181] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A resource push method, comprising: Acquire first characteristic information of interactive users corresponding to each resource type; Determining second characteristic information of each resource type according to the first characteristic information; Determining third characteristic information of resource providing users corresponding to each resource type according to a first interactive operation performed by each interactive user on a resource corresponding to the resource type; Determine, according to the fourth characteristic information of the target user, the second characteristic information and the third characteristic information, a user to be recommended corresponding to the target user, wherein there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user; The resources created by the user to be recommended are pushed to the target user.

2. The method according to claim 1, wherein: The determining, according to the first characteristic information, the second characteristic information of each resource type includes: The interactive users are clustered according to the user vectors of the interactive users of the resource type to obtain a type vector of a cluster center of the resource type, and the second feature information includes the type vector.

3. The method according to claim 2, further comprising: Determining a reliable user from among the interacting users according to the user vector and the type vector of each of the interacting users; The determining, according to the first interactive operation performed by each of the interactive users on the resource corresponding to the resource type, third characteristic information of the resource providing user corresponding to each of the resource types includes: Determining the interaction probability of each of the reliable users with respect to each of the resource providing users according to the second interaction operation performed by the reliable user with respect to the resource providing user of the resource type and the operation type of the second interaction operation; The third feature information is determined according to the user vector of the reliable user and the interaction probability.

4. The method according to claim 3, wherein: The determining, according to the fourth characteristic information of the target user, the second characteristic information, and the third characteristic information, a user to be recommended corresponding to the target user includes: Acquire, according to the fourth characteristic information and the third characteristic information, first interest information of the target user for each of the resource types; Acquire the target resource type for which the target user has a valid interactive operation; Determining a target type corresponding to the target user according to the first interest information and the target resource type; Acquire, according to the fourth characteristic information and the third characteristic information, second interest information of the target user providing the user for the resource of the target type; determining a recommendation probability based at least on the second interest information; The user to be recommended is determined according to the recommendation probability.

5. The method according to claim 4, wherein: The recommendation probability includes: the second interest information, the first interest information of the target user for the target type, and the fusion probability of the interaction of each resource providing user of the target type.

6. The method according to claim 5, wherein: The determining the recommendation probability at least according to the second interest information includes: Obtain weight configuration information, and determine type weight, hot weight and personalized weight from the weight configuration information; Obtaining a first probability according to a product of the type weight and the first interest information of the target user for the target type; Obtaining a second probability according to the product of the hot weight and the interaction probability; Obtaining a third probability according to a product of the personalized weight and the second interest information; The recommendation probability is obtained according to the sum of the first probability, the second probability and the third probability.

7. The method according to claim 4, wherein: There is no interaction operation between the target user and each resource providing user of the target type.

8. The method according to claim 2, wherein: There is a valid interaction operation between the interactive user and the resource providing user of the resource type, wherein the valid interaction operation includes an interaction operation that lasts for a valid period within a valid time period.

9. The method according to claim 3, after determining the third characteristic information of the resource providing user corresponding to each resource type according to the first interactive operation performed by each interactive user on the resource corresponding to the resource type, the method further comprises: According to the second characteristic information and the third characteristic information, from each resource providing user of the resource type, a user to be adjusted who does not belong to the resource type is obtained; Adjusting the resource type to which the user to be adjusted belongs; The corresponding interactive users, the second characteristic information and the third characteristic information are adjusted according to the adjusted resource type.

10. The method according to claim 1, wherein: The resource type is a live broadcast resource type, and the resource providing user is a live broadcast user.

11. The method according to claim 1, further comprising: Acquire the resource providing user who has interactive operations with the interactive user; The first characteristic information is generated according to the identification information of the interactive user and the identification information of each resource providing user having the interactive operation.

12. A resource push device, comprising: A first feature determination module, used to obtain first feature information of interactive users corresponding to each resource type; A second feature determination module, used to determine second feature information of each resource type according to the first feature information; A third feature determination module, configured to determine third feature information of resource providing users corresponding to each resource type according to a first interaction operation performed by each interactive user on a resource corresponding to the resource type; A recommended user determination module, used to determine a user to be recommended corresponding to the target user according to the fourth characteristic information of the target user, the second characteristic information and the third characteristic information, wherein there is no valid interaction operation between the resource type to which the user to be recommended belongs and the target user; The resource push module is used to push the resources created by the user to be recommended to the target user.

13. The device according to claim 12, wherein: The second feature determination module includes: The interactive clustering unit is used to cluster the interactive users according to the user vectors of the interactive users of the resource type to obtain the type vector of the cluster center of the resource type, and the second feature information includes the type vector.

14. The apparatus according to claim 13, further comprising: A reliable screening module, used for determining a reliable user from each of the interactive users according to the user vector and the type vector of each of the interactive users; Wherein, the third feature determination module includes: An interaction probability calculation unit, configured to determine the interaction probability of each of the reliable users with respect to each of the resource providing users according to the second interaction operation performed by the reliable user with respect to the resource providing user of the resource type and the operation type of the second interaction operation; The third feature determination unit is used to determine the third feature information according to the user vector of the reliable user and the interaction probability.

15. The device according to claim 14, wherein: The recommended user determination module includes: A first interest calculation unit, configured to obtain first interest information of the target user for each resource type according to the fourth feature information and the third feature information; An interaction type determination unit, configured to obtain a target resource type for which the target user has a valid interaction operation; a resource type screening unit, configured to determine a target type corresponding to the target user according to the first interest information and the target resource type; A second interest calculation unit, configured to obtain second interest information of the target user providing the user with respect to the resource of the target type according to the fourth feature information and the third feature information; a recommendation probability calculation unit, configured to determine a recommendation probability at least according to the second interest information; The recommended user determination unit is used to determine the user to be recommended according to the recommendation probability.

16. The device according to claim 15, wherein: The recommendation probability includes: the second interest information, the first interest information of the target user for the target type, and the fusion probability of the interaction of each resource providing user of the target type.

17. The device according to claim 16, wherein: The recommendation probability calculation unit is used to: Obtain weight configuration information, and determine type weight, hot weight and personalized weight from the weight configuration information; Calculate the product of the type weight and the first interest information of the target user for the target type to obtain a first probability; Calculate the product of the hot weight and the interaction probability to obtain a second probability; Obtaining a third probability according to a product of the personalized weight and the second interest information; The sum of the first probability, the second probability and the third probability is calculated to obtain the recommendation probability.

18. The device according to claim 15, wherein: There is no interaction operation between the target user and each resource providing user of the target type.

19. The device according to claim 13, wherein: There is a valid interaction operation between the interactive user and the resource providing user of the resource type, wherein the valid interaction operation includes an interaction operation that lasts for a valid period within a valid time period.

20. The apparatus of claim 14, further comprising: an adjustment user screening module, configured to obtain, from among the resource providing users of the resource type, users to be adjusted that do not belong to the resource type, according to the second characteristic information and the third characteristic information, after determining the third characteristic information of the resource providing users corresponding to the resource type according to the first interaction operation performed by each of the interactive users on the resource corresponding to the resource type; A resource type adjustment module, used to adjust the resource type to which the user to be adjusted belongs; The characteristic information adjustment module is used to adjust each corresponding interactive user, each second characteristic information and the third characteristic information according to the adjusted resource type.

21. The device according to claim 12, wherein: The resource type is a live broadcast type, and the resource providing user is a live broadcast user.

22. The apparatus according to claim 12, further comprising: Provide a user acquisition module for acquiring resource providing users who have interactive operations with interactive users; The feature information generating module is used to generate the first feature information according to the identification information of the interactive user and the identification information of each resource providing user with interactive operations.

23. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the resource pushing method according to any one of claims 1 to 11.

24. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute the resource pushing method according to any one of claims 1-11.

25. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the resource pushing method according to any one of claims 1 to 11.

Images