Content recommendation method, device, electronic device and storage medium

Abstract

The present disclosure relates to a content recommendation method, device, electronic device and storage medium, the method comprising: in response to a recommended content acquisition request from a client, acquiring the location information of the user account of the client; determining a target account cluster corresponding to the user account from a plurality of preset account clusters; the plurality of account clusters are clustered based on the location information of a plurality of sample accounts, and among the plurality of account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest; acquiring the recommended content associated with the target account cluster, and sending the recommended content to the client; wherein the recommended content is determined based on the interactive data of the sample accounts in the target account cluster for viewing content. This method captures the interests of user accounts more accurately and can effectively improve the accuracy of the recommended content.

Description

Technical Field

[0001] The present disclosure relates to the technical field of recommendation systems, and in particular to a content recommendation method and device, an electronic device, a storage medium, and a computer program product. Background Art

[0002] The main goal of the recommendation system is to recommend a large number of objects to users who may like them. The cold start problem is one of the important challenges faced by the recommendation system. The cold start of the recommendation system mainly includes three categories: user cold start, item cold start and system cold start. Among them, the user cold start problem refers to how to recommend objects to newly registered users to satisfy them. In the cold start problem, the system accumulates too little data for new users and often cannot make personalized recommendations to new users, resulting in the objects recommended to new users not meeting their needs.

[0003] In the prior art, the processing method for user cold start scenarios is mostly to push popular rankings, and then switch to personalized recommendations when a certain amount of user data is collected.

[0004] However, none of these recommendation methods can accurately recommend personalized content to new users. Summary of the invention

[0005] The present disclosure provides a content recommendation method, device, electronic device and readable storage medium to at least solve the problem that the cold start recommendation method in the related art cannot accurately recommend personalized content to new users. The technical solution of the present disclosure is as follows:

[0006] According to a first aspect of an embodiment of the present disclosure, a content recommendation method is provided, including:

[0007] In response to a recommended content acquisition request from a client, acquiring location information of a user account of the client;

[0008] Determining a target account cluster corresponding to the user account from a plurality of preset account clusters; the plurality of account clusters are obtained by clustering based on location information of a plurality of sample accounts, and among the plurality of account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest;

[0009] Acquire recommended content associated with the target account cluster, and send the recommended content to the client; wherein the recommended content is determined based on interaction data of sample accounts in the target account cluster for viewing content, and the interaction data is generated by interaction operations of the sample accounts for the viewing content.

[0010] In an exemplary embodiment, obtaining the recommended content associated with the target account cluster includes:

[0011] Querying cache data corresponding to the target account cluster, wherein the cache data includes a cluster center and a recommended content type corresponding to the target account cluster; the recommended content type is determined based on interaction data of sample accounts in the target account cluster for viewing content;

[0012] The target recommended content type corresponding to the target account cluster is obtained from the cache data, and the recommended content under the target recommended content type is obtained as the recommended content associated with the target account cluster.

[0013] In an exemplary embodiment, the process of determining the recommended content type corresponding to any of the account clusters includes:

[0014] Obtaining interaction data of sample accounts in the account cluster for each viewing content;

[0015] Based on the interaction data, the interaction amount of each viewing content is counted, and a plurality of target viewing contents are determined from each viewing content according to the interaction amount; the value of the interaction amount of the target viewing content is greater than the value of the interaction amount of other viewing contents, and the other viewing contents are viewing contents other than the target viewing contents in each viewing content;

[0016] The content type to which each of the target viewing contents belongs is obtained, and at least one obtained content type is used as a recommended content type corresponding to the account cluster.

[0017] In an exemplary embodiment, the obtaining of recommended content under the target recommended content type includes:

[0018] Based on the interaction volume of each viewing content under the target recommended content type, the recommended content under the target recommended content type is determined; the numerical value of the interaction volume of the recommended content is greater than the numerical value of the interaction volume of the remaining viewing content under the target recommended content type; the remaining viewing content is the viewing content other than the recommended content among the various viewing contents under the target recommended content type.

[0019] In an exemplary embodiment, the process of determining the cluster center corresponding to the account cluster includes:

[0020] The mean of the location information of the multiple sample accounts in the account cluster is obtained as the cluster center corresponding to the account cluster.

[0021] In an exemplary embodiment, the process of determining the plurality of account clusters includes:

[0022] Obtain multiple accounts with corresponding interaction data and location information as multiple sample accounts;

[0023] Acquire clustering parameters; the clustering parameters are used to determine the minimum radius of each account cluster and the minimum number of sample accounts in the account cluster;

[0024] The location information of the multiple sample accounts is clustered according to the clustering parameters, and the multiple sample accounts are clustered into multiple account clusters.

[0025] In an exemplary embodiment, the clustering parameters include multiple groups of clustering parameters, and the method further includes:

[0026] Using the multiple groups of clustering parameters to cluster the location information of the multiple sample accounts respectively, to obtain multiple groups of clustering results;

[0027] Obtaining a clustering evaluation value for each group of clustering results; the clustering evaluation value represents the clustering quality of each account cluster obtained by clustering;

[0028] The clustering result with the largest clustering evaluation value is obtained from the multiple groups of clustering results to obtain the multiple account clustering clusters.

[0029] In an exemplary embodiment, obtaining the clustering evaluation value of each group of clustering results includes:

[0030] For each group of clustering results, the number of clusters of the account clustering clusters in the group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points are obtained; the noise points represent sample accounts that do not belong to any account clustering cluster;

[0031] A weighted sum is performed on the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points to obtain a clustering evaluation value of the group of clustering results.

[0032] In an exemplary embodiment, it further includes:

[0033] When the target cluster is not determined within a preset time, obtaining preset recommended content; the preset recommended content includes at least one of the following: content determined based on historical interaction volume, content matching the registration information of the user account, and content published by an account associated with the user account;

[0034] The preset recommended content is sent to the client.

[0035] According to a second aspect of an embodiment of the present disclosure, there is provided a content recommendation device, including:

[0036] an acquisition unit, configured to execute, in response to a recommended content acquisition request from a client, acquiring location information of a user account of the client;

[0037] a determining unit configured to determine a target account cluster corresponding to the user account from a plurality of preset account clusters; the plurality of account clusters are obtained by clustering based on location information of a plurality of sample accounts, and among the plurality of account clusters, a distance between a cluster center corresponding to the target account cluster and a location corresponding to the location information of the user account is shortest;

[0038] The recommendation unit is configured to execute the acquisition of recommended content associated with the target account cluster and send the recommended content to the client; wherein the recommended content is determined based on the interaction data of the sample accounts in the target account cluster for the viewed content, and the interaction data is generated by the interaction operations of the sample accounts for the viewed content.

[0039] In an exemplary embodiment, the recommendation unit is further configured to execute a query on cached data corresponding to the target account clustering cluster, wherein the cached data includes a cluster center and a recommended content type corresponding to the target account clustering cluster; the recommended content type is determined based on interaction data of sample accounts in the target account clustering cluster for viewing content; the target recommended content type corresponding to the target account clustering cluster is obtained from the cached data, and the recommended content under the target recommended content type is obtained as the recommended content associated with the target account clustering cluster.

[0040] In an exemplary embodiment, the recommendation unit is further configured to execute the steps of obtaining interaction data of sample accounts in the account clustering cluster for each viewing content; counting the interaction volume of each viewing content based on the interaction data, and determining a plurality of target viewing contents from each viewing content according to the interaction volume; the interaction volume of the target viewing content is greater than the interaction volume of other viewing contents, and the other viewing contents are viewing contents other than the target viewing content in each viewing content; obtaining the content type to which each of the target viewing contents belongs, and using at least one obtained content type as the recommended content type corresponding to the account clustering cluster.

[0041] In an exemplary embodiment, the recommendation unit is further configured to determine the recommended content under the target recommended content type based on the interaction volume of each viewing content under the target recommended content type; the interaction volume of the recommended content is greater than the interaction volume of the remaining viewing content under the target recommended content type; the remaining viewing content is the viewing content other than the recommended content among the viewing contents under the target recommended content type.

[0042] In an exemplary embodiment, the apparatus further includes a cluster center acquisition unit configured to acquire a mean of location information of a plurality of sample accounts in the account cluster as a cluster center corresponding to the account cluster.

[0043] In an exemplary embodiment, the above-mentioned device also includes a clustering unit, which is configured to execute the acquisition of multiple accounts corresponding to interaction data and location information as multiple sample accounts; obtain clustering parameters; the clustering parameters are used to determine the minimum radius of each account clustering cluster and the minimum number of sample accounts in the account clustering cluster; cluster the location information of the multiple sample accounts according to the clustering parameters, and cluster the multiple sample accounts into multiple account clustering clusters.

[0044] In an exemplary embodiment, the clustering unit is further configured to execute clustering of the location information of the multiple sample accounts respectively using the multiple groups of clustering parameters to obtain multiple groups of clustering results; obtain a clustering evaluation value for each group of clustering results; the clustering evaluation value represents the clustering quality of each account clustering cluster obtained by clustering; obtain the clustering result with the largest clustering evaluation value from the multiple groups of clustering results to obtain the multiple account clustering clusters.

[0045] In an exemplary embodiment, the clustering unit is further configured to execute, for each group of clustering results, obtaining the number of clusters of the account clustering clusters in the group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points; the noise points represent sample accounts that do not belong to any account clustering cluster; and performing weighted summation on the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points to obtain a clustering evaluation value for the group of clustering results.

[0046] In an exemplary embodiment, the recommendation unit is further configured to obtain preset recommended content when the target cluster is not determined within a preset time; the preset recommended content includes at least one of the following: content determined based on historical interaction volume, content matching the registration information of the user account, and content published by an associated account of the user account; and the preset recommended content is sent to the client.

[0047] According to a third aspect of an embodiment of the present disclosure, there is provided a content recommendation electronic device, including:

[0048] processor;

[0049] a memory for storing instructions executable by the processor;

[0050] The processor is configured to execute the instructions to implement the content recommendation method as described in any one of the embodiments of the first aspect.

[0051] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, comprising: when instructions in the computer-readable storage medium are executed by a processor of a content recommendation electronic device, the content recommendation electronic device is enabled to execute the content recommendation method described in any one of the embodiments of the first aspect.

[0052] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, the computer program product comprising a computer program, the computer program being stored in a computer-readable storage medium, at least one processor of a device reading and executing the computer program from the computer-readable storage medium, so that the device executes the content recommendation method described in any one of the embodiments of the first aspect.

[0053] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0054] By pre-stating the location information of sample accounts and the interactive data for viewing content, and clustering multiple sample accounts based on the location information, multiple account clusters are obtained. By counting the interactive data of sample accounts in each account cluster, and determining the recommended content associated with each account cluster, when receiving a client's recommended content acquisition request, the location information of the user account can be obtained, and the target cluster corresponding to the user account can be determined from multiple account clusters based on the location information of the user account, and the recommended content associated with the obtained target cluster can be pushed to the user account. This method constructs an association relationship between group interests and individual interests, so that even for new user accounts without any historical viewing behavior, it can be mapped to the corresponding group based on the location information of the new user account, and the content of interest to the corresponding group can be recommended to the new user account. Compared with the conventional cold start recommendation method, this method captures the interests of user accounts more accurately and can effectively improve the accuracy of the recommended content.

[0055] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0057] Figure 1 The present invention is a flowchart of a content recommendation method according to an exemplary embodiment.

[0058] Figure 2 The figure is a flowchart of steps for determining the recommended content type corresponding to an account cluster according to an exemplary embodiment.

[0059] Figure 3 The figure is a flowchart of another content recommendation method according to an exemplary embodiment.

[0060] Figure 4 The present invention is a flowchart of model construction and iterative optimization of a content recommendation method according to an exemplary embodiment.

[0061] Figure 5 The figure is a structural block diagram of a content recommendation device according to an exemplary embodiment.

[0062] Figure 6 It is a block diagram of a device according to an exemplary embodiment. DETAILED DESCRIPTION

[0063] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0064] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. "Multiple" in the specification and claims of the present disclosure and the drawings refers to two or more. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0065] The content recommendation method provided by the present disclosure can be applied to a server, a terminal, or a system including a terminal and a server, and implemented through the interaction between the terminal and the server. The terminal can be, but is not limited to, various smart phones, tablet computers, or laptop computers, and the server can be implemented as an independent server or a server cluster composed of multiple servers.

[0066] Figure 1 is a flow chart of a content recommendation method according to an exemplary embodiment. Figure 1 As shown, the content recommendation method is applied in the server, comprising the following steps:

[0067] In step S102, in response to the recommended content acquisition request of the client, the location information of the user account of the client is acquired.

[0068] The recommended content acquisition request may be a request for acquiring recommended content for an application installed in the client.

[0069] The triggering of the recommended content acquisition request may be triggered when the application is detected to be started, or may be triggered after the application is started through a trigger entry preset in the application.

[0070] The location information may be the latitude and longitude information of the user account obtained through LBS (Location Based Services).

[0071] The user account may be a new user account that uses the application platform for the first time, or a historical user account that has used the application platform. The following description will take the user account as a new user account as an example to illustrate the present disclosure.

[0072] In a specific implementation, when a new user account needs to obtain recommended content of an application installed in the client, the client may be triggered to generate a recommended content acquisition request, which is sent by the client to the server. The server receives and responds to the recommended content acquisition request to obtain the location information of the user account of the client.

[0073] In step S104, a target account cluster corresponding to the user account is determined from a plurality of preset account clusters; the plurality of account clusters are obtained by clustering based on location information of a plurality of sample accounts, and among the plurality of account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest.

[0074] The sample accounts are accounts that have interactive data and location information for viewing content, and each account cluster contains multiple sample accounts.

[0075] The cluster center may represent the center position of the account cluster.

[0076] In a specific implementation, the server may pre-count the location information of sample accounts within a certain period, and then use a clustering algorithm to cluster the sample accounts based on the location information of each sample account to obtain multiple account clusters, and may pre-count the cluster centers of each account cluster. Thus, after obtaining the location information of a new user account, the target cluster that is most relevant to the new user account may be determined based on the location information of the new user account and the cluster centers of each cluster.

[0077] More specifically, based on the location information of the new user account and the cluster centers of each cluster cluster, the process of determining the target cluster cluster may include: based on the location information of the new user account and the cluster centers of each cluster cluster, respectively obtaining the relative distances between the cluster centers of each account cluster cluster and the position corresponding to the location information of the new user account, determining the relative distance with the smallest value from each relative distance, and using the cluster center corresponding to the relative distance with the smallest value as the target cluster center, and using the cluster cluster corresponding to the target cluster center as the target cluster cluster most relevant to the new user account.

[0078] In step S106, the recommended content associated with the target account cluster is obtained and sent to the client; wherein the recommended content is determined based on the interactive data of the sample accounts in the target account cluster for the viewed content, and the interactive data is generated by the interactive operations of the sample accounts for the viewed content.

[0079] Among them, the recommended content can represent the group interest information of each account cluster.

[0080] Among them, the interactive data can represent the interest information of the sample account in the viewed content. For example, the interactive data may include behavioral data such as viewing time, comments, likes, collections, sharing and forwarding of the viewed content.

[0081] In a specific implementation, after obtaining multiple account clusters, the group interest information of each account cluster can also be counted. Specifically, the interaction volume corresponding to each viewing content can be counted based on the interaction data of the sample accounts in each account cluster for the viewing content, and the interaction volume can be sorted according to the numerical value. The viewing content corresponding to the top-ranked interaction volume is used as the group interest content of each account cluster, and then as the recommended content associated with each account cluster. Therefore, after determining the target account cluster corresponding to the new user account, the recommended content associated with the target account cluster can be obtained, and the recommended content can be sent to the client of the new user account.

[0082] In the above content recommendation method, the location information of sample accounts and the interactive data for viewing content are pre-counted, and multiple sample accounts are clustered based on the location information to obtain multiple account clustering clusters. The interactive data of sample accounts in each account clustering cluster is counted, and the recommended content associated with each account clustering cluster is determined. When a client request for obtaining recommended content is received later, the location information of the user account can be obtained, and the target clustering cluster corresponding to the user account can be determined from multiple account clustering clusters based on the location information of the user account, and the recommended content associated with the obtained target clustering cluster can be pushed to the user account. This method constructs an association relationship between group interests and individual interests, so that even for new user accounts without any historical viewing behavior, it can be mapped to the corresponding group based on the location information of the new user account, and the content of interest to the corresponding group can be recommended to the new user account. Compared with the conventional cold start recommendation method, this method captures the interests of user accounts more accurately and can effectively improve the accuracy of the recommended content.

[0083] In an exemplary embodiment, the step of obtaining recommended content associated with the target account cluster in step S106 specifically includes: querying cache data corresponding to the target account cluster, the cache data including the cluster center and recommended content type corresponding to the target account cluster; the recommended content type is determined based on the interaction data of sample accounts in the target account cluster for viewing content; obtaining the target recommended content type corresponding to the target account cluster from the cache data, and obtaining the recommended content under the target recommended content type as the recommended content associated with the target account cluster.

[0084] Among them, the recommended content type can represent the common points of the recommended content. For example, the recommended content type can be maternal and child type, fitness type, film and television type, entertainment type, etc.

[0085] In a specific implementation, when the server counts the group interest information of each account clustering cluster, it can first determine, for each account clustering cluster, the multiple viewing contents for which the sample accounts in the account clustering cluster have interactive data, and then count the interactive volume of each viewing content, and use the content types of the multiple viewing contents with the highest interactive volume as the recommended content types of the account clustering cluster, and further upload the obtained recommended content types of each account clustering cluster to the cache database. Thus, after determining the target account clustering cluster corresponding to the user account, the cache data corresponding to the target account clustering cluster can be queried from the cache database, and the target recommended content type corresponding to the target account clustering cluster can be determined from the cache data. Furthermore, after determining the target content type, the interactive volume of the viewing content currently under the target recommended content type can be obtained, and based on the interactive volume of the viewing content currently under the target recommended content type, the target recommended content is determined, and pushed to the user account as the recommended content associated with the target account clustering cluster.

[0086] It is understandable that in another embodiment, the interactive data of multiple sample accounts for the viewing content under the target recommended content type can also be counted in advance, and the multiple viewing contents ranked in the top order by the obtained interactive volume are used as the recommended content under the target recommended content type, and the target recommended content type and the recommended content are stored in another server accordingly. After the target recommended content type is determined, the recommended content corresponding to the target recommended content is obtained from the server and returned to the client of the user account. Thus, by storing the recommended content type and the account cluster in the cache database, and storing the recommended content type and the recommended content under the recommended content type in another server, the resource usage of the cache database can be saved.

[0087] The technical solution provided in this embodiment determines the recommended content type through the interactive data of the sample accounts in the target account clustering cluster for the viewed content, and stores it in the cache database, so that after determining the target account clustering cluster, the target recommended content type associated with the target account clustering cluster can be determined based on the cache data in the cache database, and the recommended content currently under the target recommended content type can be obtained and pushed to the user account, so that the user account can obtain the real-time recommended content under the target recommended content type associated with the corresponding target clustering cluster, thereby achieving accurate determination of the recommended content for the user account and improving the real-time nature of the obtained recommended content.

[0088] In an exemplary embodiment, if Figure 2 As shown, the process of determining the recommended content type corresponding to any account cluster includes:

[0089] Step S202, obtaining the interaction data of the sample accounts in the account cluster for each viewing content;

[0090] Step S204, counting the interaction amount of each viewing content based on the interaction data, and determining a plurality of target viewing contents from each viewing content according to the interaction amount; the value of the interaction amount of the target viewing content is greater than the value of the interaction amount of other viewing contents, and the other viewing contents are viewing contents other than the target viewing contents in each viewing content;

[0091] Step S206, obtaining the content type to which each target viewing content belongs, and using at least one obtained content type as a recommended content type corresponding to the account cluster.

[0092] Among them, the interaction volume can represent the total amount of interaction data such as likes, shares and reposts for the viewed content.

[0093] In a specific implementation, since the interactive data can characterize the interest information of the sample accounts for the viewing content, the group interest information corresponding to the account clustering cluster can be determined based on the interactive data of each sample account in the account clustering cluster for the viewing content, and the viewing content with more interactive data, that is, the viewing content with a larger interactive volume, is taken as the group interest of the account clustering cluster. More specifically, the interactive data such as likes, shares and forwarding of each viewing content by the sample accounts in the account clustering cluster can be obtained first, and then for each viewing content, the total amount of interactive data of all sample accounts in the account clustering cluster for the viewing content is counted as the interactive volume of each viewing content, and the interactive volume of each viewing content is further sorted in order of numerical value, and the first multiple viewing contents with larger numerical values ​​are taken as target viewing contents, for example, the viewing contents with the top 5 interactive volumes are taken as target viewing contents, and finally, the content type to which each target viewing content belongs is obtained, and at least one obtained content type is taken as the recommended content type corresponding to the account clustering cluster.

[0094] More specifically, in one embodiment, since the interactive data for each viewed content may include multiple behavioral data such as likes, shares, comments, favorites, and forwarding, the method for determining the interactive volume of each viewed content may be: pre-set corresponding weight coefficients for different types of interactive data, separately count the interactive volumes of different types of interactive data for each viewed content, and perform weighted summation of the interactive volumes of different types of interactive data using the corresponding weight coefficients to obtain the interactive volume of each viewed content.

[0095] The technical solution provided by this embodiment determines the interaction volume of each viewing content through the interaction data of sample accounts in the account clustering cluster for each viewing content, and then determines multiple target viewing contents ranked with the highest interaction volume based on the interaction volume, and uses the content type to which the target viewing content belongs as the recommended content type corresponding to the account clustering cluster. This method uses the content type to which the viewing content with a large interaction volume in each account clustering cluster belongs as the recommended content type of the group interest of the account clustering cluster, so that the obtained recommended content type has high universality, thereby improving the matching between the recommended content pushed to the user account and the user account.

[0096] In an exemplary embodiment, the specific steps of obtaining recommended content under the target recommended content type include: determining the recommended content under the target recommended content type based on the interaction volume of each viewing content under the target recommended content type; the interaction volume of the recommended content is greater than the interaction volume of the remaining viewing content under the target recommended content type; the remaining viewing content is the viewing content other than the recommended content among the viewing content under the target recommended content type.

[0097] In a specific implementation, after determining the target recommended content type corresponding to the target account cluster, the interaction data of the currently viewed content under the target recommended content type can be counted to obtain the interaction volume of each viewed content under the target recommended content type, and the interaction volume can be sorted according to the numerical value. The first few viewed contents with larger interaction volume values, for example, the first 3 viewed contents, can be pushed to the user account as recommended content under the target recommended content type.

[0098] The technical solution provided by this embodiment makes the first multiple viewing contents with large interaction volume under the target recommended content type as the recommended contents under the target recommended content type, so that the obtained recommended contents have high universality, thereby improving the matching of the recommended contents with the user account.

[0099] In an exemplary embodiment, the process of determining the cluster center corresponding to the account cluster includes: obtaining the mean value of the location information of multiple sample accounts in the account cluster as the cluster center corresponding to the account cluster.

[0100] In this embodiment, the mean of the location information of each sample account in the account clustering cluster, that is, the central position of the account clustering cluster, is used as the clustering center corresponding to the account clustering cluster. This facilitates improving the reliability and accuracy of the determination result when determining the target account clustering cluster corresponding to the user account based on the acquired user account location information and clustering center.

[0101] In an exemplary embodiment, the process of determining multiple account clustering clusters includes: obtaining multiple accounts corresponding to interaction data and location information as multiple sample accounts; obtaining clustering parameters; the clustering parameters are used to determine the minimum radius of each account clustering cluster and the minimum number of sample accounts in the account clustering cluster; clustering the location information of the multiple sample accounts according to the clustering parameters, and clustering the multiple sample accounts into multiple account clustering clusters.

[0102] The clustering parameters may include the minimum radius of the account cluster and the minimum number of sample accounts included in the account cluster.

[0103] Among them, the account cluster is the largest set of density-connected points.

[0104] In a specific implementation, multiple accounts that have interactive data and location information for viewing content within a certain period can be obtained as multiple sample accounts, and a density-based clustering algorithm is used to cluster the location information of multiple sample accounts in combination with the minimum radius of the account clustering cluster and the minimum number of sample accounts included in the account clustering cluster to obtain multiple account clustering clusters. More specifically, a circle (called a neighborhood) is drawn with each sample account as the center and the minimum radius as the radius, and the number of sample accounts in the circle is obtained as the density value of the center point of the circle. If the center point of the circle where the number of sample accounts in the circle is less than the minimum number of sample accounts included in the account clustering cluster is recorded as a low-density point, and the center point of the circle where the number of sample accounts in the circle is greater than or equal to the minimum number of sample accounts included in the account clustering cluster is recorded as a high-density point. If a high-density point is within the circle of another high-density point, the two points can be connected, thereby connecting multiple points continuously in series. Afterwards, if a low-density point is also within the circle of a high-density point, the low-density point is connected to the nearest high-density point, which is called a boundary point. Therefore, all the points that can be connected together constitute an account cluster, and the low-density points that are not in the circle of any high-density points are noise points.

[0105] The technical solution provided in this embodiment clusters multiple sample accounts into multiple account clustering clusters through a density-based clustering algorithm, can find clusters of various sizes and shapes, and has certain anti-noise characteristics, thereby improving the clustering effect.

[0106] In an exemplary embodiment, the clustering parameters include multiple groups of clustering parameters, and also include: using the multiple groups of clustering parameters to cluster the location information of multiple sample accounts respectively to obtain multiple groups of clustering results; obtaining a clustering evaluation value for each group of clustering results; the clustering evaluation value represents the clustering quality of each account clustering cluster obtained by clustering; obtaining the clustering result with the largest clustering evaluation value from the multiple groups of clustering results to obtain multiple account clustering clusters.

[0107] The clustering evaluation value is used to characterize the clustering quality of the account clustering clusters obtained by clustering. The larger the clustering evaluation value is, the better the clustering result is, and the smaller the clustering evaluation value is, the worse the clustering result is.

[0108] In the specific implementation, the minimum radius of the account clustering cluster and the minimum number of sample accounts included in the account clustering cluster are adjustable clustering parameters. In order to further improve the clustering effect, multiple groups of clustering parameters can be set, and each group of clustering parameters is used to cluster the location information of multiple sample accounts to obtain multiple groups of clustering effects. For each group of clustering results, the clustering evaluation value of each group of clustering results is calculated according to the preset clustering evaluation value calculation method, and the numerical values ​​of the clustering evaluation values ​​of each group of clustering results are compared to determine the clustering evaluation value with the largest numerical value. The multiple account clustering clusters in the clustering results corresponding to the clustering evaluation value with the largest numerical value are used as the final account clustering clusters obtained after clustering the sample accounts.

[0109] Furthermore, in an exemplary embodiment, the step of obtaining a clustering evaluation value for each group of clustering results includes: for each group of clustering results, obtaining the number of clusters of the account clustering clusters in the group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points; the noise points represent sample accounts that do not belong to any account clustering cluster; and performing weighted summation on the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points to obtain the clustering evaluation value of the group of clustering results.

[0110] In a specific implementation, for each group of clustering results, the number of clusters of the account clusters in the clustering results, the cumulative sum of the number of user accounts contained in each account cluster, and the number of noise points can be obtained respectively, and the weight coefficients corresponding to the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points are obtained respectively, and the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points are weighted and summed by their corresponding weight coefficients respectively to obtain the clustering evaluation value of each group of clustering results. Among them, the weight coefficient represents the degree of influence of the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points on the clustering effect.

[0111] If the cumulative sum of the number of clusters, the number of user accounts, and the number of noise points are recorded as X1, X2, and X3, respectively, and the corresponding weight coefficients are recorded as w1, w2, and w3, respectively, the relationship between the cluster evaluation value (represented as R) can be expressed as:

[0112] R=X1*w1+X2*w2+X3*w3

[0113] It is understandable that after obtaining multiple groups of clustering results through multiple groups of clustering parameters, each group of clustering results can also be displayed as a visualized geographical location rendering diagram, and the final clustering result can be determined from each group of clustering results through the visualized geographical location rendering diagram.

[0114] The technical solution provided in this embodiment obtains multiple groups of clustering results through multiple groups of clustering parameters, determines the clustering evaluation value of each group of clustering results through the number of clusters of account clustering clusters in each group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points, and selects the clustering result with the best clustering effect from each group of clustering results based on the clustering evaluation value, thereby improving the clustering effect of the obtained multiple account clustering clusters.

[0115] In an exemplary embodiment, it also includes: when the target cluster is not determined within a preset time, obtaining preset recommended content; the preset recommended content includes at least one of the following: content determined based on historical interaction volume, content matching the registration information of the user account, and content published by an associated account of the user account; and sending the preset recommended content to the client.

[0116] The technical solution provided in this embodiment takes into account the abnormal situation when the target cluster is not determined within the preset time, and proposes a fallback strategy for obtaining preset recommended content and sending it to the client of the user account, which can avoid the situation where the user account cannot obtain the recommended content.

[0117] Figure 3 is a flowchart of another content recommendation method according to an exemplary embodiment. Figure 3 As shown, in this embodiment, the method includes the following steps:

[0118] Step S302, in response to the recommended content acquisition request from the client, acquiring location information of the user account of the client;

[0119] Step S304, determining a target account cluster corresponding to the user account from a plurality of preset account clusters; the plurality of account clusters are clustered based on the location information of the plurality of sample accounts, and among the plurality of account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest;

[0120] Step S306, querying cache data corresponding to the target account cluster, the cache data including the cluster center and recommended content type corresponding to the target account cluster; the recommended content type is determined based on the interactive data of the sample accounts in the target account cluster for viewing the content;

[0121] Step S308, obtaining the target recommended content type corresponding to the target account cluster from the cache data, and obtaining the recommended content under the target recommended content type as the recommended content associated with the target account cluster.

[0122] The content recommendation method provided in this embodiment obtains multiple account clusters by pre-stating the location information of sample accounts and the interactive data for viewing content, and clustering multiple sample accounts based on the location information. The interactive data of sample accounts in each account cluster is counted, and the recommended content associated with each account cluster is determined. When a client request for obtaining recommended content is received, the location information of the user account is obtained, and the target cluster corresponding to the user account is determined from multiple account clusters based on the location information of the user account, and the recommended content associated with the obtained target cluster is pushed to the user account. This method constructs an association relationship between group interests and individual interests, so that even for new user accounts without any historical viewing behavior, it can be mapped to the corresponding group based on the location information of the new user account, and the content of interest to the corresponding group can be recommended to the new user account. Compared with the conventional cold start recommendation method, this method captures the interests of the user account more accurately, and can effectively improve the accuracy of the recommended content.

[0123] Figure 4 The present invention is a flowchart of model construction and iterative optimization of a content recommendation method according to an exemplary embodiment, comprising the following steps:

[0124] (1) Pre-collect statistics on user accounts that have clicked or completed playback within a certain period, as well as the LBS information (latitude and longitude information) of each user account, and the video category data of the user account that has clicked or completed playback;

[0125] (2) Clustering the LBS information of all user accounts using a density-based clustering algorithm, where the density-based clustering algorithm can be a density-based clustering method with noise (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), which divides areas with sufficient density into clusters and finds clusters of arbitrary shapes in a spatial database with noise. It defines a cluster as the largest set of density-connected points. The adjustable parameters in the clustering process include: the minimum radius of the clustering cluster and the minimum number of sample points contained in the clustering cluster. The evaluation result is optimized by adjusting the parameter combination.

[0126] (3) Count the videos clicked by the user accounts of each cluster obtained in step (2), and statistically analyze the TOP5 categories of each cluster video. At the same time, average the LBS data of the cluster to obtain the center LBS coordinates of each cluster. The LBS coordinates of the center point of each cluster and the TOP5 categories corresponding to the cluster are stored in the online cache device for online query. The online cache data format is:

[0127] <key:value>=<latitude and longitude coordinates, List<category 1, category 2, category 3>>

[0128] (4) After the online user account enters the platform, the platform retrieves the longitude and latitude center point closest to the user account from the cache database obtained in step (3) that stores the longitude and latitude of the center point of the cluster and the TOP categories of clicked videos corresponding to the cluster based on the longitude and latitude information of the user account in the traffic request parameters, and returns the category corresponding to the mapped center point. Then, based on the returned category, the TOP videos clicked under the category are returned as recommended videos.

[0129] (5) In step (4), a fallback strategy is set. When the cluster corresponding to the user account is not found within a preset time, a preset recommended video, such as a popular video on the entire network, can be returned to the user account.

[0130] The above content recommendation method obtains multiple account clusters by pre-stating the location information of sample accounts and the interactive data for viewing content, and clustering multiple sample accounts based on the location information. The interactive data of sample accounts in each account cluster is counted, and the recommended content associated with each account cluster is determined. When receiving a request for obtaining recommended content from a client, the location information of the user account can be obtained, and the target cluster corresponding to the user account can be determined from multiple account clusters based on the location information of the user account, and the recommended content associated with the obtained target cluster can be pushed to the user account. This method constructs an association relationship between group interests and individual interests, so that even for new user accounts without any historical viewing behavior, it can be mapped to the corresponding group based on the location information of the new user account, and the content of interest to the corresponding group can be recommended to the new user account. Compared with conventional cold start recommendation methods, this method captures the interests of user accounts more accurately and can effectively improve the accuracy of recommended content.

[0131] It should be understood that although Figure 1-4 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1-4 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0132] Figure 5 is a structural block diagram of a content recommendation device according to an exemplary embodiment. Figure 5 The device includes an acquisition unit 502, a determination unit 504 and a recommendation unit 506, wherein:

[0133] The acquisition unit 502 is configured to execute the response to the recommended content acquisition request of the client and acquire the location information of the user account of the client;

[0134] The determining unit 504 is configured to determine a target account cluster corresponding to the user account from a plurality of preset account clusters; the plurality of account clusters are obtained by clustering based on the location information of the plurality of sample accounts, and among the plurality of account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest;

[0135] The recommendation unit 506 is configured to execute the acquisition of recommended content associated with the target account cluster and send the recommended content to the client; wherein the recommended content is determined based on the interaction data of the sample accounts in the target account cluster for the viewed content, and the interaction data is generated by the interactive operations of the sample accounts for the viewed content.

[0136] In an exemplary embodiment, the recommendation unit 506 is further configured to execute a query on cached data corresponding to the target account clustering cluster, wherein the cached data includes a cluster center and a recommended content type corresponding to the target account clustering cluster; the recommended content type is determined based on the interaction data of sample accounts in the target account clustering cluster for viewing content; the target recommended content type corresponding to the target account clustering cluster is obtained from the cached data, and the recommended content under the target recommended content type is obtained as the recommended content associated with the target account clustering cluster.

[0137] In an exemplary embodiment, the recommendation unit 506 is further configured to execute the steps of obtaining interaction data of sample accounts in the account clustering cluster for each viewing content; counting the interaction amount of each viewing content based on the interaction data, and determining multiple target viewing content from each viewing content according to the interaction amount; the interaction amount of the target viewing content is greater than the interaction amount of other viewing content, and the other viewing content is viewing content other than the target viewing content in each viewing content; obtaining the content type to which each target viewing content belongs, and using at least one obtained content type as the recommended content type corresponding to the account clustering cluster.

[0138] In an exemplary embodiment, the recommendation unit 506 is further configured to determine the recommended content under the target recommended content type based on the interaction amount of each viewing content under the target recommended content type; the numerical value of the interaction amount of the recommended content is greater than the numerical value of the interaction amount of the remaining viewing content under the target recommended content type; the remaining viewing content is the viewing content other than the recommended content among the various viewing contents under the target recommended content type.

[0139] In an exemplary embodiment, the apparatus further includes a cluster center acquisition unit configured to acquire a mean of location information of a plurality of sample accounts in an account cluster as a cluster center corresponding to the account cluster.

[0140] In an exemplary embodiment, the above-mentioned device also includes a clustering unit, which is configured to execute the acquisition of multiple accounts corresponding to interaction data and location information as multiple sample accounts; obtain clustering parameters; the clustering parameters are used to determine the minimum radius of each account clustering cluster and the minimum number of sample accounts in the account clustering cluster; cluster the location information of multiple sample accounts according to the clustering parameters, and cluster the multiple sample accounts into multiple account clustering clusters.

[0141] In an exemplary embodiment, the clustering unit is further configured to execute clustering of location information of multiple sample accounts using multiple groups of clustering parameters to obtain multiple groups of clustering results; obtain a clustering evaluation value for each group of clustering results; the clustering evaluation value represents the clustering quality of each account clustering cluster obtained by clustering; obtain the clustering result with the largest clustering evaluation value from the multiple groups of clustering results to obtain multiple account clustering clusters.

[0142] In an exemplary embodiment, the clustering unit is further configured to execute, for each group of clustering results, obtaining the number of clusters of the account clustering clusters in the group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points; the noise points represent sample accounts that do not belong to any account clustering cluster; and performing weighted summation on the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points to obtain a clustering evaluation value for the group of clustering results.

[0143] In an exemplary embodiment, the recommendation unit 506 is further configured to execute, when the target cluster is not determined within a preset time, obtaining preset recommended content; the preset recommended content includes at least one of the following: content determined based on historical interaction volume, content matching the registration information of the user account, and content published by an associated account of the user account; and sending the preset recommended content to the client.

[0144] Regarding the device in the above embodiment, the specific manner in which each unit performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0145] Figure 6 6 is a block diagram of a device 600 for executing the above content recommendation method according to an exemplary embodiment. For example, the device 600 may be a server. Figure 6 , the device 600 includes a processing component 620, which further includes one or more processors, and a memory resource represented by a memory 622 for storing instructions that can be executed by the processing component 620, such as an application. The application stored in the memory 622 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 620 is configured to execute instructions to perform the above-mentioned data query method.

[0146] The device 600 may also include a power supply component 624 configured to perform power management of the device 600, a wired or wireless network interface 626 configured to connect the device 600 to a network, and an input / output (I / O) interface 628. The device 600 may operate based on an operating system stored in the memory 622, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or the like.

[0147] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 622 including instructions, which can be executed by a processor of the device 600 to perform the above method. For example, the computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0148] In an exemplary embodiment, a computer program product is also provided, including a computer program, which is stored in a computer-readable storage medium, and at least one processor of a device reads and executes the computer program from the computer-readable storage medium, so that the device performs the content recommendation method described in any embodiment of the present disclosure.

[0149] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0150] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.< / key:value>

Claims

1. A content recommendation method, characterized in that: include: In response to a recommended content acquisition request from a client, acquiring location information of a user account of the client; The user account of the client includes a new user account; Determining a target account cluster corresponding to the user account from a plurality of preset account clusters; The multiple account clusters are obtained by clustering based on the location information of the multiple sample accounts, and among the multiple account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest; Acquire the recommended content associated with the target account cluster, and send the recommended content to the client; wherein the recommended content is determined based on the interactive data of the sample accounts in the target account cluster for the viewed content, and the interactive data is generated by the interactive operation of the sample accounts for the viewed content; The process of determining the multiple account clusters includes: Using multiple groups of clustering parameters to cluster the location information of the multiple sample accounts respectively, to obtain multiple groups of clustering results; For each group of clustering results, the number of clusters of the account clustering clusters in the group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points are obtained; the noise points represent sample accounts that do not belong to any account clustering cluster; Performing weighted summation on the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points to obtain a clustering evaluation value of the group of clustering results; The clustering result with the largest clustering evaluation value is obtained from each group of clustering results to obtain the multiple account clustering clusters.

2. The method according to claim 1, characterized in that The obtaining of the recommended content associated with the target account cluster includes: Querying cache data corresponding to the target account cluster, wherein the cache data includes a cluster center and a recommended content type corresponding to the target account cluster; the recommended content type is determined based on interaction data of sample accounts in the target account cluster for viewing content; The target recommended content type corresponding to the target account cluster is obtained from the cache data, and the recommended content under the target recommended content type is obtained as the recommended content associated with the target account cluster.

3. The method according to claim 2, characterized in that The process of determining the recommended content type corresponding to any of the account clusters includes: Obtaining interaction data of sample accounts in the account cluster for each viewing content; Based on the interaction data, the interaction amount of each viewing content is counted, and a plurality of target viewing contents are determined from each viewing content according to the interaction amount; the value of the interaction amount of the target viewing content is greater than the value of the interaction amount of other viewing contents, and the other viewing contents are viewing contents other than the target viewing contents in each viewing content; The content type to which each of the target viewing contents belongs is obtained, and at least one obtained content type is used as a recommended content type corresponding to the account cluster.

4. The method according to any one of claims 2 to 3, characterized in that: The obtaining of recommended content under the target recommended content type includes: Based on the interaction volume of each viewing content under the target recommended content type, the recommended content under the target recommended content type is determined; the numerical value of the interaction volume of the recommended content is greater than the numerical value of the interaction volume of the remaining viewing content under the target recommended content type; the remaining viewing content is the viewing content other than the recommended content among the various viewing contents under the target recommended content type.

5. The method according to claim 1, characterized in that The process of determining the cluster center corresponding to the account cluster includes: The mean of the location information of the multiple sample accounts in the account cluster is obtained as the cluster center corresponding to the account cluster.

6. The method according to claim 1, characterized in that The process of determining the multiple account clusters includes: Obtain multiple accounts with corresponding interaction data and location information as multiple sample accounts; Acquire clustering parameters; the clustering parameters are used to determine the minimum radius of each account cluster and the minimum number of sample accounts in the account cluster; The location information of the multiple sample accounts is clustered according to the clustering parameters, and the multiple sample accounts are clustered into multiple account clusters.

7. The method according to claim 1, characterized in that Also includes: When the target account cluster is not determined within a preset time, obtaining preset recommended content; The preset recommended content includes at least one of the following: content determined based on historical interaction volume, content matching the registration information of the user account, and content published by an account associated with the user account; The preset recommended content is sent to the client.

8. A content recommendation device, characterized in that: include: an acquisition unit, configured to execute, in response to a recommended content acquisition request from a client, acquiring location information of a user account of the client; The user account of the client includes a new user account; a determining unit configured to determine a target account cluster corresponding to the user account from a plurality of preset account clusters; The multiple account clusters are obtained by clustering based on the location information of the multiple sample accounts, and among the multiple account clusters, the distance between the cluster center corresponding to the target account cluster and the location corresponding to the location information of the user account is the shortest; A recommendation unit, configured to obtain recommended content associated with the target account cluster, and send the recommended content to the client; wherein the recommended content is determined based on the interactive data of the sample accounts in the target account cluster for the viewed content, and the interactive data is generated by the interactive operation of the sample accounts for the viewed content; The clustering unit is configured to execute clustering of the location information of the multiple sample accounts respectively using multiple groups of clustering parameters to obtain multiple groups of clustering results; for each group of clustering results, obtain the number of clusters of the account clustering clusters in the group of clustering results, the cumulative sum of the number of user accounts included in each account clustering cluster, and the number of noise points; the noise points represent sample accounts that do not belong to any account clustering cluster; perform weighted summation on the number of clusters, the cumulative sum of the number of user accounts, and the number of noise points to obtain a clustering evaluation value of the group of clustering results; obtain the clustering result with the largest clustering evaluation value from each group of clustering results to obtain the multiple account clustering clusters.

9. The device according to claim 8, characterized in that The recommendation unit is further configured to execute a query on cache data corresponding to the target account clustering cluster, wherein the cache data includes a cluster center and a recommended content type corresponding to the target account clustering cluster; the recommended content type is determined based on interaction data of sample accounts in the target account clustering cluster for viewing content; the target recommended content type corresponding to the target account clustering cluster is obtained from the cache data, and the recommended content under the target recommended content type is obtained as the recommended content associated with the target account clustering cluster.

10. The device according to claim 9, characterized in that The recommendation unit is further configured to execute the steps of obtaining interaction data of sample accounts in the account clustering cluster for each viewing content; counting the interaction amount of each viewing content based on the interaction data, and determining a plurality of target viewing contents from each viewing content according to the interaction amount; the interaction amount of the target viewing content is greater than the interaction amount of other viewing contents, and the other viewing contents are viewing contents other than the target viewing contents in each viewing content; obtaining the content type to which each of the target viewing contents belongs, and using at least one of the obtained content types as the recommended content type corresponding to the account clustering cluster.

11. The device according to any one of claims 9 to 10, characterized in that: The recommendation unit is further configured to determine the recommended content under the target recommended content type based on the interaction amount of each viewing content under the target recommended content type; the value of the interaction amount of the recommended content is greater than the value of the interaction amount of the remaining viewing content under the target recommended content type; the remaining viewing content is the viewing content other than the recommended content among the viewing contents under the target recommended content type.

12. The device according to claim 8, characterized in that The above-mentioned device also includes a cluster center acquisition unit, which is configured to execute the acquisition of the mean value of the location information of multiple sample accounts in the account cluster cluster as the cluster center corresponding to the account cluster cluster.

13. The device according to claim 8, characterized in that The clustering unit is further configured to execute the acquisition of multiple accounts corresponding to interaction data and location information as multiple sample accounts; acquisition of clustering parameters; the clustering parameters are used to determine the minimum radius of each account clustering cluster and the minimum number of sample accounts in the account clustering cluster; clustering the location information of the multiple sample accounts according to the clustering parameters, and clustering the multiple sample accounts into multiple account clustering clusters.

14. The device according to claim 8, characterized in that The recommendation unit is further configured to obtain preset recommended content when the target account cluster is not determined within a preset time; The preset recommended content includes at least one of the following: content determined based on historical interaction volume, content matching the registration information of the user account, and content published by an account associated with the user account; the preset recommended content is sent to the client.

15. A content recommendation electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the content recommendation method as described in any one of claims 1 to 7.

16. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of a content recommendation electronic device, the content recommendation electronic device is enabled to perform the content recommendation method as claimed in any one of claims 1 to 7.

Images