Resource delivery methods, devices, computer equipment and media

Abstract

This disclosure relates to a resource push method, apparatus, computer device, and medium, belonging to the field of Internet technology. In the embodiments of this disclosure, when pushing resources to a target object, weight coefficients corresponding to each behavioral dimension are determined based on the first and second parameters of the target object in each behavioral dimension. Thus, in determining the weight coefficients for each behavioral dimension, not only are the differences of the target object relative to the object group in each behavioral dimension considered, but also the distribution of multiple candidate resources of the target object in each behavioral dimension. This improves both the efficiency and accuracy of determining the weight coefficients. Furthermore, based on the behavioral prediction parameters of the candidate resources of the target object in each behavioral dimension and the corresponding weight coefficients, push parameters for the candidate resources are determined, and resource push is then performed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

Description

Technical Field

[0001] This disclosure relates to the field of Internet technology, and in particular to a resource push method, apparatus, computer equipment, and medium. Background Technology

[0002] With the rapid development of internet technology and the gradual expansion of the internet user base, internet-based multimedia resources have permeated all aspects of people's lives. Simultaneously, methods for pushing multimedia resources have also developed rapidly. Typically, when pushing multimedia resources to internet users, various types of interactive behaviors that users have exhibited with other multimedia resources—such as positive interactions like clicking, liking, and saving, and negative interactions like blocking and skipping—as well as the viewing time of other multimedia resources, all affect the results of the multimedia resource push. Therefore, the multimedia resource push problem is a multi-objective (i.e., multi-behavioral dimension) optimization problem. The optimization goal is to maximize positive user interactions while ensuring the viewing time of multimedia resources.

[0003] Currently, when facing the above-mentioned multi-objective optimization problem, it is necessary to balance the relationship between various behavioral dimensions. Usually, business personnel set a weight coefficient for each behavioral dimension based on historical experience, and then determine the push parameters of multimedia resources based on the push sub-parameters of each behavioral dimension and the corresponding weight coefficients of each behavioral dimension. Multimedia resources are then pushed to network users based on the push parameters of multimedia resources.

[0004] However, the above process relies on human determination, which reduces the efficiency and accuracy of resource delivery. Summary of the Invention

[0005] This disclosure provides a resource push method, apparatus, computer device, and medium, which improves both the efficiency and accuracy of determining weight coefficients, thereby enhancing the efficiency and accuracy of resource push. The technical solution of this disclosure is as follows:

[0006] According to a first aspect of the present disclosure, a resource push method is provided, the method comprising:

[0007] Obtain behavioral prediction parameters of candidate resources of the target object in at least two behavioral dimensions. These behavioral prediction parameters represent the predicted interaction behavior between the target object and the candidate resource in that behavioral dimension.

[0008] Based on the first and second parameters of the target object in this behavioral dimension, the weight coefficient corresponding to this behavioral dimension is determined. The first parameter represents the difference of the target object relative to the object group in this behavioral dimension, and the second parameter represents the distribution of multiple candidate resources of the target object in this behavioral dimension.

[0009] Based on the behavior prediction parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions, the push parameters of the candidate resource are determined. These push parameters are used to determine the push order of the candidate resource when pushing resources to the target object.

[0010] In some embodiments, the process of obtaining the first parameter includes:

[0011] Obtain the first average value of the target object in the behavioral dimension and the second average value of the object group in the behavioral dimension. The first average value is the average value of the behavioral prediction parameters of the candidate resources of the target object in the behavioral dimension, and the second average value is the average value of the behavioral prediction parameters of the candidate resources of the multiple objects included in the object group in the behavioral dimension.

[0012] Based on the first average value and the second average value, the first parameter is determined, wherein the first parameter is positively correlated with the first average value and negatively correlated with the second average value.

[0013] In some embodiments, the process of obtaining the second parameter includes:

[0014] Obtain the dispersion of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determine the dispersion as the second parameter.

[0015] In some embodiments, determining the weight coefficient corresponding to the behavioral dimension based on the first and second parameters of the target object on that behavioral dimension includes:

[0016] Based on the first and second parameters of the target object in this behavioral dimension, a third parameter of the target object in this behavioral dimension is determined, wherein the third parameter is positively correlated with the first parameter and positively correlated with the second parameter;

[0017] Based on the initial weight coefficient corresponding to the behavioral dimension and the third parameter, the weight coefficient corresponding to the behavioral dimension is determined, wherein the weight coefficient is positively correlated with the initial weight coefficient and the third parameter.

[0018] In some embodiments, determining the third parameter of the target object in the behavioral dimension based on the first and second parameters of the target object includes:

[0019] Based on the second parameter of the target object in this behavioral dimension, a fourth parameter of the target object in this behavioral dimension is determined, and the fourth parameter is negatively correlated with the second parameter;

[0020] Based on the first and fourth parameters of the target object in this behavioral dimension, the third parameter is determined, and the third parameter is negatively correlated with the fourth parameter.

[0021] In some embodiments, obtaining behavioral prediction parameters of candidate resources of a target object across at least two behavioral dimensions includes:

[0022] Based on the object feature information of the target object and the content feature information of the candidate resource, the behavioral prediction parameters of the candidate resource in at least two behavioral dimensions are determined.

[0023] In some embodiments, determining the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions based on the object feature information of the target object and the content feature information of the candidate resource includes:

[0024] The object feature information of the target object and the content feature information of the candidate resource are input into the behavior prediction model corresponding to the behavior dimension. The behavior prediction model predicts the behavior prediction parameters of the candidate resource in the behavior dimension, and the behavior prediction parameters of the candidate resource in the behavior dimension are obtained. The behavior prediction model is trained based on the object feature information of the sample object, the content feature information of the candidate resource of the sample object, and the behavior parameter labels of the candidate resource of the sample object in the behavior dimension.

[0025] In some embodiments, determining the push parameters of the candidate resource based on the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions includes:

[0026] On at least two behavioral dimensions, the candidate resources are sorted according to the behavioral prediction parameters on at least two behavioral dimensions.

[0027] Based on the ranking order of the candidate resource in at least two behavioral dimensions, push sub-parameters of the candidate resource in at least two behavioral dimensions are determined respectively, and the push sub-parameters are negatively correlated with the ranking order;

[0028] The push parameters of the candidate resource are determined based on the push sub-parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0029] In some embodiments, sorting the candidate resources according to the behavioral prediction parameters of the at least two behavioral dimensions includes:

[0030] In the queues corresponding to the at least two behavioral dimensions, the candidate resources are sorted according to the behavioral prediction parameters of the at least two behavioral dimensions.

[0031] In some embodiments, determining the push parameters of a candidate resource based on the push sub-parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions includes:

[0032] The push parameters of the candidate resource are obtained by weighted summation based on the push sub-parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0033] According to a second aspect of the present disclosure, a resource delivery device is provided, the device comprising:

[0034] The acquisition unit is configured to acquire behavior prediction parameters of candidate resources of the target object in at least two behavioral dimensions, the behavior prediction parameters representing the predicted interaction behavior between the target object and the candidate resource in the behavioral dimension;

[0035] The weight coefficient determination unit is configured to perform a first parameter and a second parameter on the target object in the behavioral dimension to determine the weight coefficient corresponding to the behavioral dimension. The first parameter represents the difference of the target object relative to the object group in the behavioral dimension, and the second parameter represents the distribution of multiple candidate resources of the target object in the behavioral dimension.

[0036] The push parameter determination unit is configured to execute a push parameter determination for the candidate resource based on the behavior prediction parameters of the candidate resource in at least two behavior dimensions and the weight coefficients corresponding to the at least two behavior dimensions. The push parameter is used to determine the push order of the candidate resource when pushing resources to the target object.

[0037] In some embodiments, the weighting coefficient determining unit includes:

[0038] The first acquisition subunit is configured to acquire the first average value of the target object in the behavioral dimension and the second average value of the object group in the behavioral dimension. The first average value is the average value of the behavioral prediction parameters of the candidate resources of the target object in the behavioral dimension, and the second average value is the average value of the behavioral prediction parameters of the candidate resources of the multiple objects included in the object group in the behavioral dimension.

[0039] The determination subunit is configured to perform the determination of the first parameter based on the first average value and the second average value, wherein the first parameter is positively correlated with the first average value and negatively correlated with the second average value.

[0040] In some embodiments, the weighting coefficient determining unit includes:

[0041] The second acquisition subunit is configured to acquire the dispersion of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determine the dispersion as the second parameter.

[0042] In some embodiments, the weighting coefficient determining unit includes:

[0043] The third parameter determination subunit is configured to execute the first and second parameters of the target object in the behavioral dimension to determine the third parameter of the target object in the behavioral dimension, wherein the third parameter is positively correlated with the first parameter and the second parameter.

[0044] The weight coefficient determination subunit is configured to perform the determination of the weight coefficient corresponding to the behavior dimension based on the initial weight coefficient and the third parameter, wherein the weight coefficient is positively correlated with the initial weight coefficient and the third parameter.

[0045] In some embodiments, the third parameter determines the subunit, which is configured to perform:

[0046] Based on the second parameter of the target object in this behavioral dimension, a fourth parameter of the target object in this behavioral dimension is determined, and the fourth parameter is negatively correlated with the second parameter;

[0047] Based on the first and fourth parameters of the target object in this behavioral dimension, the third parameter is determined, and the third parameter is negatively correlated with the fourth parameter.

[0048] In some embodiments, the acquisition unit is configured to perform:

[0049] Based on the object feature information of the target object and the content feature information of the candidate resource, the behavioral prediction parameters of the candidate resource in at least two behavioral dimensions are determined.

[0050] In some embodiments, the acquisition unit is configured to perform:

[0051] The object feature information of the target object and the content feature information of the candidate resource are input into the behavior prediction model corresponding to the behavior dimension. The behavior prediction model predicts the behavior prediction parameters of the candidate resource in the behavior dimension, and the behavior prediction parameters of the candidate resource in the behavior dimension are obtained. The behavior prediction model is trained based on the object feature information of the sample object, the content feature information of the candidate resource of the sample object, and the behavior parameter labels of the candidate resource of the sample object in the behavior dimension.

[0052] In some embodiments, the push parameter determining unit includes:

[0053] The sorting subunit is configured to sort the candidate resource according to the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions, respectively, in the at least two behavioral dimensions.

[0054] The sub-unit is configured to perform a push sub-parameter for the candidate resource in the at least two behavioral dimensions based on the ranking order of the candidate resource in the at least two behavioral dimensions, wherein the push sub-parameter is negatively correlated with the ranking order.

[0055] The determined subunit is further configured to perform push parameters for the candidate resource based on the push sub-parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0056] In some embodiments, the sorting subunit is configured to perform:

[0057] In the queues corresponding to the at least two behavioral dimensions, the candidate resources are sorted according to the behavioral prediction parameters of the at least two behavioral dimensions.

[0058] In some embodiments, the determining subunit is further configured to perform:

[0059] The push parameters of the candidate resource are obtained by weighted summation based on the push sub-parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0060] According to a third aspect of the present disclosure, a computer device is provided, the computer device comprising:

[0061] One or more processors;

[0062] Memory used to store the executable program code of the processor;

[0063] The processor is configured to execute the program code to implement the resource push method described above.

[0064] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, the computer-readable storage medium comprising: when program code in the computer-readable storage medium is executed by a processor of a computer device, enabling the computer device to perform the resource push method described above.

[0065] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the resource push method described above.

[0066] The technical solution provided in this disclosure, when pushing resources to a target object, determines the weight coefficients corresponding to each behavioral dimension based on the first and second parameters of the target object in each behavioral dimension. Thus, in determining the weight coefficients for each behavioral dimension, not only are the differences of the target object relative to the object group in each behavioral dimension considered, but also the distribution of multiple candidate resources of the target object in each behavioral dimension. This ensures that the determined weight coefficients not only reflect the tendency of the target object in each behavioral dimension, but also the distribution characteristics of the candidate resources in each behavioral dimension. For different objects, the first and second parameters of each object in different behavioral dimensions are referenced to determine the weight coefficients for each object in different behavioral dimensions, enabling the determination of different weight coefficients for different objects. This results in personalized weight coefficients that reflect the object characteristics and resource characteristics of each object, improving both the efficiency and accuracy of weight coefficient determination. Furthermore, based on the behavioral prediction parameters of the target object's candidate resources in each behavioral dimension and the corresponding weight coefficients, the push parameters for the candidate resources are determined, and resources are then pushed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

[0067] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0068] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0069] Figure 1 This is a schematic diagram illustrating the implementation environment of a resource push method according to an exemplary embodiment;

[0070] Figure 2 This is a flowchart illustrating a resource push method according to an exemplary embodiment;

[0071] Figure 3 This is a flowchart illustrating a resource push method according to an exemplary embodiment;

[0072] Figure 4 This is a flowchart illustrating a resource push method according to an exemplary embodiment;

[0073] Figure 5 This is a block diagram illustrating a resource push device according to an exemplary embodiment;

[0074] Figure 6This is a block diagram illustrating a terminal according to an exemplary embodiment;

[0075] Figure 7 This is a block diagram illustrating a server according to an exemplary embodiment. Detailed Implementation

[0076] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0077] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0078] The data or information disclosed herein may be data or information authorized by the user or fully authorized by all parties.

[0079] First, the application scenarios involved in the embodiments of this disclosure will be described:

[0080] The resource push method provided in this disclosure can be applied to multimedia resource push scenarios, such as short video push scenarios. In some embodiments, when facing multi-objective (i.e., multi-behavioral dimension) optimization problems in push scenarios, the resource push method provided in this disclosure can be used to balance the relationship between various behavioral dimensions, such as balancing the relationship between the viewing time of multimedia resources and positive interaction behavior.

[0081] It should be noted that the process of pushing multimedia resources typically involves three stages: recall, coarse ranking, and fine ranking. In some embodiments, the resource pushing method provided in this disclosure is applied to the fine ranking stage of resource pushing.

[0082] The push notification process consists of three phases: The recall phase, the first phase, involves quickly retrieving a subset of potentially interesting multimedia resources from a massive multimedia resource library based on user and resource characteristics, and adding these resources to the recall queue. The coarse ranking phase, the second phase, sorts the multimedia resources in the recall queue according to coarse ranking metrics (such as click-through rate) to filter out the top-ranked resources, adding them to the coarse ranking queue to reduce the number of recalled resources and alleviate the ranking pressure in the fine ranking phase. The fine ranking phase, the third phase, further sorts the multimedia resources in the coarse ranking queue according to fine ranking metrics (such as revenue per thousand impressions) to create a fine ranking queue, selecting the top-ranked multimedia resources for push notification.

[0083] Figure 1 This is a schematic diagram illustrating the implementation environment of a resource push method provided in this embodiment of the disclosure. See also... Figure 1 The implementation environment includes: server 101.

[0084] Server 101 can be a standalone physical server, a server cluster consisting of multiple physical servers, or a distributed file system. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0085] In some embodiments, server 101 provides a function to push multimedia resources to users. In this embodiment, server 101 is used to obtain behavioral prediction parameters of candidate resources for a target object in at least two behavioral dimensions; based on the first and second parameters of the target object in those behavioral dimensions, determine the weight coefficient corresponding to that behavioral dimension; and based on the behavioral prediction parameters of the candidate resource in those at least two behavioral dimensions and the weight coefficients corresponding to those at least two behavioral dimensions, determine the push parameters of the candidate resource. In this embodiment, "target object" will be used to refer to the user to whom multimedia resources are to be pushed.

[0086] In some embodiments, the number of servers 101 described above may be more or fewer, and this disclosure does not limit this. Of course, servers 101 may also include other functional servers to provide more comprehensive and diversified services.

[0087] In some embodiments, the implementation environment further includes a terminal 102. In some embodiments, the server 101 and the terminal 102 are directly or indirectly connected via wired or wireless communication, which is not limited in this disclosure.

[0088] Terminal 102 can be at least one of the following devices: smartphone, smartwatch, desktop computer, laptop, virtual reality terminal, augmented reality terminal, wireless terminal, and laptop computer. Terminal 102 has communication capabilities and can access wired or wireless networks. Terminal 102 can refer to one of multiple terminals; this embodiment uses terminal 102 as an example only. Those skilled in the art will understand that the number of terminals can be more or less.

[0089] In some embodiments, terminal 102 runs an application with multimedia resource playback capabilities, such as a social application, a live streaming application, a short video application, etc. In some embodiments, server 101 provides background services for the application running on terminal 102. In this embodiment of the present disclosure, terminal 102 is used to send a resource push request to server 101 in response to a resource push request triggered by a target object, so as to trigger server 101 to push multimedia resources to the target object.

[0090] Figure 2 This is a flowchart illustrating a resource push method according to an exemplary embodiment, such as... Figure 2 As shown, this method is executed by the server and includes the following steps:

[0091] In step 201, the server obtains the behavior prediction parameters of the candidate resources of the target object in at least two behavior dimensions. The behavior prediction parameters represent the predicted behavior of the target object interacting with the candidate resources in the corresponding behavior dimensions.

[0092] In step 202, the server determines the weight coefficient corresponding to the behavioral dimension based on the first parameter and the second parameter of the target object in the behavioral dimension. The first parameter represents the difference of the target object relative to the object group in the behavioral dimension, and the second parameter represents the distribution of the multiple candidate resources of the target object in the behavioral dimension.

[0093] In step 203, the server determines the push parameters of the candidate resource based on the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions. The push parameters are used to determine the push order of the candidate resource when pushing resources to the target object.

[0094] The technical solution provided in this disclosure, when pushing resources to a target object, determines the weight coefficients corresponding to each behavioral dimension based on the first and second parameters of the target object in each behavioral dimension. Thus, in determining the weight coefficients for each behavioral dimension, not only are the differences of the target object relative to the object group in each behavioral dimension considered, but also the distribution of multiple candidate resources of the target object in each behavioral dimension. This ensures that the determined weight coefficients not only reflect the tendency of the target object in each behavioral dimension, but also the distribution characteristics of the candidate resources in each behavioral dimension. For different objects, the first and second parameters of each object in different behavioral dimensions are referenced to determine the weight coefficients for each object in different behavioral dimensions, enabling the determination of different weight coefficients for different objects. This results in personalized weight coefficients that reflect the object characteristics and resource characteristics of each object, improving both the efficiency and accuracy of weight coefficient determination. Furthermore, based on the behavioral prediction parameters of the target object's candidate resources in each behavioral dimension and the corresponding weight coefficients, the push parameters for the candidate resources are determined, and resources are then pushed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

[0095] In some embodiments, the process of obtaining the first parameter includes:

[0096] Obtain the first average value of the target object in the behavioral dimension and the second average value of the object group in the behavioral dimension. The first average value is the average value of the behavioral prediction parameters of the candidate resources of the target object in the behavioral dimension, and the second average value is the average value of the behavioral prediction parameters of the candidate resources of the multiple objects included in the object group in the behavioral dimension.

[0097] Based on the first average value and the second average value, the first parameter is determined, wherein the first parameter is positively correlated with the first average value and negatively correlated with the second average value.

[0098] In some embodiments, the process of obtaining the second parameter includes:

[0099] Obtain the dispersion of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determine the dispersion as the second parameter.

[0100] In some embodiments, determining the weight coefficient corresponding to the behavioral dimension based on the first and second parameters of the target object on that behavioral dimension includes:

[0101] Based on the first and second parameters of the target object in this behavioral dimension, a third parameter of the target object in this behavioral dimension is determined, wherein the third parameter is positively correlated with the first parameter and positively correlated with the second parameter;

[0102] Based on the initial weight coefficient corresponding to the behavioral dimension and the third parameter, the weight coefficient corresponding to the behavioral dimension is determined, wherein the weight coefficient is positively correlated with the initial weight coefficient and the third parameter.

[0103] In some embodiments, determining the third parameter of the target object in the behavioral dimension based on the first and second parameters of the target object includes:

[0104] Based on the second parameter of the target object in this behavioral dimension, a fourth parameter of the target object in this behavioral dimension is determined, and the fourth parameter is negatively correlated with the second parameter;

[0105] Based on the first and fourth parameters of the target object in this behavioral dimension, the third parameter is determined, and the third parameter is negatively correlated with the fourth parameter.

[0106] In some embodiments, obtaining behavioral prediction parameters of candidate resources of a target object across at least two behavioral dimensions includes:

[0107] Based on the object feature information of the target object and the content feature information of the candidate resource, the behavioral prediction parameters of the candidate resource in at least two behavioral dimensions are determined.

[0108] In some embodiments, determining the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions based on the object feature information of the target object and the content feature information of the candidate resource includes:

[0109] The object feature information of the target object and the content feature information of the candidate resource are input into the behavior prediction model corresponding to the behavior dimension. The behavior prediction model predicts the behavior prediction parameters of the candidate resource in the behavior dimension, and the behavior prediction parameters of the candidate resource in the behavior dimension are obtained. The behavior prediction model is trained based on the object feature information of the sample object, the content feature information of the candidate resource of the sample object, and the behavior parameter labels of the candidate resource of the sample object in the behavior dimension.

[0110] In some embodiments, determining the push parameters of the candidate resource based on the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions includes:

[0111] On at least two behavioral dimensions, the candidate resources are sorted according to the behavioral prediction parameters on at least two behavioral dimensions.

[0112] Based on the ranking order of the candidate resource in at least two behavioral dimensions, push sub-parameters of the candidate resource in at least two behavioral dimensions are determined respectively, and the push sub-parameters are negatively correlated with the ranking order;

[0113] The push parameters of the candidate resource are determined based on the push sub-parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0114] In some embodiments, sorting the candidate resources according to the behavioral prediction parameters of the at least two behavioral dimensions includes:

[0115] In the queues corresponding to the at least two behavioral dimensions, the candidate resources are sorted according to the behavioral prediction parameters of the at least two behavioral dimensions.

[0116] In some embodiments, determining the push parameters of a candidate resource based on the push sub-parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions includes:

[0117] The push parameters of the candidate resource are obtained by weighted summation based on the push sub-parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0118] The above Figure 2 The diagram shown is merely a basic flow of this disclosure. The following description, based on a specific implementation method, further elaborates on the solution provided by this disclosure. Figure 3 This is a flowchart illustrating a resource push method according to an exemplary embodiment, see [link / reference]. Figure 3 The method includes:

[0119] In step 301, the terminal sends a resource push request for the target object to the server. This resource push request is used to request the server to push resources for the target object.

[0120] Here, the terminal refers to the terminal corresponding to the target object. The target object refers to the target user, and the pushed resource is the pushed multimedia resource; the target user is the user to whom the multimedia resource is to be pushed. In some embodiments, the target object is represented by an object identifier, such as an object name, object account, or object ID (Identity).

[0121] In this embodiment of the disclosure, the terminal runs an application with multimedia resource playback capabilities, such as a social networking application, a live streaming application, a short video application, etc. Correspondingly, the server is the backend server of the aforementioned application with multimedia resource playback capabilities. In some embodiments, the multimedia resource is any one of video, image, or article. In some embodiments, the multimedia resource is represented by a resource identifier, for example, a resource name, resource number, resource ID, etc.

[0122] In some embodiments, the resource push request is triggered based on the target object's launch operation of the aforementioned application with multimedia resource playback function. For example, if the target object operates on the terminal and launches the aforementioned application with multimedia resource playback function, the terminal responds to the target object's launch operation of the application and sends a resource push request to the server to request the server to push multimedia resources for the target object. Alternatively, the resource push request is triggered based on the target object's resource switching operation in the application. For example, if the target object switches between currently displayed multimedia resources while watching multimedia resources in the application, the terminal responds to the target object's switching operation of the multimedia resources and sends a resource push request to the server to request the server to push multimedia resources for the target object. Of course, the resource push request can also be triggered based on other operations of the target object, such as the target object's access operation of the push page in the application, the target object's refresh operation in the application, or the target object's search operation in the application, etc., and this disclosure does not limit this.

[0123] In step 302, in response to the resource push request of the target object, the server obtains the behavior prediction parameters of the candidate resources of the target object in at least two behavioral dimensions. The behavior prediction parameters represent the prediction of the interaction behavior between the target object and the candidate resources in the behavioral dimension.

[0124] The resource push request carries the object identifier of the target object. In some embodiments, after receiving the resource push request for the target object, the server obtains the object identifier of the target object carried in the resource push request, obtains candidate resources for the target object based on the object identifier, and then obtains behavioral prediction parameters of the candidate resources in at least two behavioral dimensions based on the candidate resources of the target object.

[0125] In this embodiment of the disclosure, candidate resources refer to candidate multimedia resources. In some embodiments, the number of candidate resources is multiple. In some embodiments, candidate resources refer to multiple multimedia resources included in the coarse-ranking queue, and then, in the fine-ranking stage of the push, the subsequent resource push process is performed based on the multiple multimedia resources included in the coarse-ranking queue and in conjunction with the resource push method provided in this embodiment of the disclosure.

[0126] In this embodiment of the disclosure, the behavioral dimension refers to the dimension corresponding to different types of interactive behaviors. In some embodiments, the behavioral dimension includes at least two of the following dimensions: click, like, follow, comment, enter profile, favorite, block, skip, and viewing duration of multimedia resources. Accordingly, the behavioral prediction parameters include at least two of the following: click rate, like rate, follow rate, comment rate, enter profile rate, favorite rate, block rate, skip rate, and predicted viewing duration. Among them, click-through rate (CTR) represents the probability that a target audience clicks on the candidate resource; like rate represents the probability that a target audience likes the candidate resource; follow rate represents the probability that a target audience follows the candidate resource; comment rate represents the probability that a target audience comments on the candidate resource; personal page visit rate represents the probability that a target audience visits the personal page associated with the candidate resource; favorite rate represents the probability that a target audience favorites the candidate resource; block rate represents the probability that a target audience blocks the candidate resource; skip rate represents the probability that a target audience skips the candidate resource; and predicted viewing time represents the predicted viewing time of the candidate resource. The above embodiments illustrate various types of behavioral dimensions. Of course, in other embodiments, the behavioral dimensions also include other types of behavioral dimensions, such as forwarding dimensions, the number of times multimedia resources are viewed, etc., and this disclosure does not limit these aspects.

[0127] In some embodiments, the server determines the behavior prediction parameters of the candidate resource in at least two behavioral dimensions based on the object feature information of the target object and the content feature information of the candidate resource.

[0128] The object feature information represents the object characteristics of the target object. In some embodiments, the object feature information includes the target object's attribute feature information and behavioral feature information. The attribute feature information represents the target object's attribute characteristics, such as gender, age, city, hobbies, etc., while the behavioral feature information represents the target object's behavioral characteristics, such as clicking, searching, liking, and collecting. In some embodiments, the server obtains the above attribute feature information and behavioral feature information based on the target object's user profile, which describes the object's attribute and behavioral information. The content feature information represents the content characteristics of the candidate resource. In some embodiments, the content feature information is the text feature or image feature of the candidate resource. In this embodiment, using the target object's object feature information and the candidate resource's content feature information to predict the interaction behavior between the target object and the candidate resource can improve the prediction accuracy of the behavior prediction parameters.

[0129] In some embodiments, the server employs a behavior prediction model to predict the behavior prediction parameters of the candidate resource across at least two behavior dimensions. The process is as follows: for any behavior dimension, the server inputs the object feature information of the target object and the content feature information of the candidate resource into the behavior prediction model corresponding to that behavior dimension. The behavior prediction model then predicts the behavior prediction parameters of the candidate resource across that behavior dimension, thus obtaining the behavior prediction parameters of the candidate resource across that behavior dimension. In this way, by utilizing the behavior prediction models corresponding to each behavior dimension to predict the interaction behavior between the target object and the candidate resource, both the prediction efficiency and accuracy of the behavior prediction parameters are improved.

[0130] In this embodiment of the disclosure, the behavior prediction model is a pre-trained model. In some embodiments, for any behavior dimension, the behavior prediction model is trained based on the object feature information of the sample object, the content feature information of the candidate resources of the sample object, and the behavior parameter labels of the candidate resources of the sample object in that behavior dimension. The behavior parameter labels represent the interaction behavior between the sample object and the candidate resource in that behavior dimension.

[0131] In some embodiments, for any behavior dimension, the server obtains object feature information of multiple sample objects, content feature information of candidate resources of the multiple sample objects, and behavior parameter labels of the candidate resources of the multiple sample objects in that behavior dimension. Based on the object feature information of the multiple sample objects, the content feature information of the candidate resources of the multiple sample objects, and the behavior parameter labels of the candidate resources of the multiple sample objects in that behavior dimension, the server trains a model to obtain a behavior prediction model. Specifically, the training process of the behavior prediction model includes: in the first iteration, the object feature information of the multiple sample objects and the content feature information of the candidate resources of the multiple sample objects are input into the initial model to obtain the parameter training results of the first iteration; based on the parameter training results of the first iteration and the behavior parameter labels of the corresponding sample objects, a loss function is determined; based on the loss function, the model parameters in the initial model are adjusted; the model parameters adjusted in the first iteration are used as the model parameters for the second iteration, and the second iteration is performed; the above iteration process is repeated multiple times. In the Nth iteration, the model parameters adjusted in the (N-1)th iteration are used as the new model parameters for model training until the training meets the target conditions. The model corresponding to the iteration process that meets the target conditions is then obtained as the behavior prediction model. Wherein, N is a positive integer, and N is greater than 1. In some embodiments, the target condition for training is that the number of training iterations of the initial model reaches a target number, which is a pre-set number of training iterations; or, the target condition for training is that the loss value meets a target threshold condition, such as the loss value being less than 0.00001. The embodiments of this disclosure do not limit the setting of the target condition.

[0132] In step 303, for any behavioral dimension, the server obtains the first parameter of the target object in that behavioral dimension, which represents the difference of the target object relative to the object group in that behavioral dimension.

[0133] Here, "object group" refers to the user group. In some embodiments, "object group" refers to all objects using the aforementioned application with multimedia resource playback function, or, "object group" refers to objects that have logged into the aforementioned application with multimedia resource playback function.

[0134] In some embodiments, the server obtains a first average value of the target object in the behavioral dimension and a second average value of the object group in the behavioral dimension, and determines the first parameter based on the first average value and the second average value. The first average value is the average of the behavioral prediction parameters of the candidate resources of the target object in the behavioral dimension, and the second average value is the average of the behavioral prediction parameters of the candidate resources of the multiple objects included in the object group in the behavioral dimension.

[0135] In this embodiment, the first parameter is positively correlated with the first average value and negatively correlated with the second average value. Accordingly, the process by which the server determines the first parameter based on the first and second average values ​​is as follows: In one optional embodiment, the server determines the ratio of the first average value to the second average value and uses this ratio as the first parameter; or, in another optional embodiment, the server determines the difference between the first and second average values ​​and uses this difference as the first parameter. It should be noted that the above-described method of determining the ratio as the first parameter or the difference as the first parameter is an exemplary method of this embodiment. In other embodiments, the server may use other calculation methods that can represent the correlation between the first parameter and the first and second average values ​​to determine the first parameter, and this embodiment does not limit this. The following embodiments of this disclosure will use the example of determining the ratio as the first parameter to illustrate the scheme.

[0136] In some embodiments, before implementing this solution, the server determines the behavior prediction parameters of the candidate resources of the multiple objects included in the object group on each behavior dimension. Optionally, the server periodically determines the behavior prediction parameters of the candidate resources of the multiple objects included in the object group on each behavior dimension and updates the behavior prediction parameters of the candidate resources of the multiple objects included in the object group on each behavior dimension in real time.

[0137] For example, taking the target user as the target object, candidate videos as candidate resources, and likes as the behavioral dimension, if the number of candidate videos for the target user is 10, and assuming the like rates of these 10 candidate videos are a1, a2, ..., a9, a10 respectively, then the first average value is (a1+a2+...+a9+a10) / 10; if the number of candidate videos for the target group is 100, and assuming the like rates of these 100 candidate videos are b1, b2, ..., b99, b100 respectively, then the second average value is (b1+b2+...+b99+b100) / 100; and thus, the first parameter value is the ratio of (a1+a2+...+a9+a10) / 10 to (b1+b2+...+b99+b100) / 100.

[0138] In the above embodiments, the first parameter is determined based on the first average value and the second average value. Since the first average value is the average value of the behavior prediction parameters of the target object in the behavior dimension, and the second average value is the average value of the behavior prediction parameters of the multiple objects included in the object group in the behavior dimension, the determined first parameter can reflect the difference between the target object and the object group in the behavior dimension, so that the weight coefficient corresponding to the behavior dimension can be determined based on the first parameter.

[0139] In step 304, the server obtains a second parameter of the target object in the behavioral dimension, which represents the distribution of multiple candidate resources of the target object in the behavioral dimension.

[0140] In some embodiments, the server obtains the dispersion of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determines the dispersion as the second parameter.

[0141] Here, dispersion is used to represent the degree of dispersion of a dataset. Accordingly, the process by which the server determines the second parameter is as follows: In one optional embodiment, the server obtains the variance of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determines the variance as the second parameter; or, in another optional embodiment, the server obtains the standard deviation of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determines the standard deviation as the second parameter. It should be noted that the above-described method of determining the variance or standard deviation as the second parameter is an exemplary method of this disclosure embodiment. In other embodiments, the server may choose other parameters that can represent the degree of dispersion as the second parameter, and this disclosure embodiment does not limit this. The following embodiments of this disclosure will use the determination of the standard deviation as the second parameter as an example to illustrate the scheme.

[0142] For example, taking the standard deviation as the second parameter, the standard deviation of the behavior prediction parameters of multiple candidate resources in the like dimension can represent the distribution of the multiple candidate resources in the like dimension; the standard deviation of the behavior prediction parameters of multiple candidate resources in the collection dimension can represent the distribution of the multiple candidate resources in the collection dimension.

[0143] In the above embodiment, the dispersion of the behavior prediction parameters of multiple candidate resources on the behavior dimension is determined as the second parameter. Since the dispersion can reflect the degree of dispersion of a dataset, the determined second parameter can reflect the distribution of multiple candidate resources on the behavior dimension, so that the weight coefficients corresponding to the behavior dimension can be determined based on the second parameter.

[0144] In step 305, the server determines the weight coefficient corresponding to the behavioral dimension based on the first and second parameters of the target object in that behavioral dimension.

[0145] In some embodiments, the server determines a third parameter of the target object in the behavioral dimension based on a first parameter and a second parameter of the target object in the behavioral dimension, wherein the third parameter is positively correlated with the first parameter and the second parameter; furthermore, the server determines a weight coefficient corresponding to the behavioral dimension based on an initial weight coefficient corresponding to the behavioral dimension and the third parameter, wherein the weight coefficient is positively correlated with the initial weight coefficient and the third parameter.

[0146] In this embodiment of the disclosure, the third parameter is positively correlated with the first parameter and positively correlated with the second parameter. The process by which the server determines the third parameter based on the first and second parameters is described below. The corresponding process is as follows: Based on the second parameter of the target object in this behavioral dimension, a fourth parameter of the target object in this behavioral dimension is determined, and this fourth parameter is negatively correlated with the second parameter; based on the first and fourth parameters of the target object in this behavioral dimension, the third parameter is determined, and this third parameter is negatively correlated with the fourth parameter.

[0147] The fourth parameter is negatively correlated with the second parameter. Accordingly, the process by which the server determines the fourth parameter based on the second parameter is as follows: In one optional embodiment, the server determines a difference between the second parameter and the target object in the behavioral dimension, and uses this difference as the fourth parameter of the target object in the behavioral dimension; or, in another optional embodiment, after determining the difference between the second parameter and the target object, the server determines the square of this difference and uses this square as the fourth parameter of the target object in the behavioral dimension. In this embodiment, by performing the above-described process of taking the difference or squaring the difference on the second parameter, the fourth parameter can be quickly determined, and the determined fourth parameter reflects the negative correlation between the second and second parameters, so that the weight coefficient corresponding to the behavioral dimension can be determined subsequently based on the fourth parameter. Of course, in other embodiments, the server can use other calculation methods that can represent the negative correlation between the fourth and second parameters to determine the fourth parameter, such as taking the reciprocal of the second parameter to obtain the fourth parameter, etc. This disclosure does not limit this. The following embodiments of this disclosure will use the example of determining the squared value as the fourth parameter to illustrate the scheme.

[0148] The third parameter is positively correlated with the first parameter and negatively correlated with the fourth parameter. Accordingly, the server determines the third parameter based on the first and fourth parameters as follows: In one optional embodiment, the server determines the ratio of the first parameter to the fourth parameter and uses this ratio as the third parameter; or, in another optional embodiment, the server determines the difference between the first and fourth parameters and uses this difference as the third parameter. It should be noted that the above-described method of determining the ratio or the difference as the third parameter is an exemplary method of this disclosure. In other embodiments, the server may use other calculation methods that can represent the correlation between the third parameter and the first and fourth parameters to determine the third parameter, and this disclosure does not limit this approach. The following embodiments of this disclosure will use the example of determining the ratio as the third parameter to illustrate the solution.

[0149] In the above embodiments, an intermediate parameter (fourth parameter) that is negatively correlated with the second parameter is determined. Then, based on the positive correlation between the third parameter and the first parameter, and the negative correlation between the third parameter and the intermediate parameter, the third parameter is determined. This ensures that the determined third parameter reflects not only the positive correlation with the first parameter but also the positive correlation with the second parameter, so that the third parameter can be used to determine the weight coefficient corresponding to the behavioral dimension. It should be noted that the above method of determining the fourth parameter first and then the third parameter is an exemplary method of this disclosure. In other embodiments, the server may use other calculation methods that can represent the positive correlation between the third parameter and the first and second parameters to determine the third parameter. For example, the server may determine the sum of the first and second parameters as the third parameter; or the server may determine the product of the first and second parameters as the third parameter, etc. This disclosure does not limit this.

[0150] In this embodiment, the weight coefficient is positively correlated with the initial weight coefficient and positively correlated with the third parameter. The process by which the server determines the weight coefficient corresponding to the behavior dimension based on the initial weight coefficient and the third parameter is described below. The process is as follows: In one optional embodiment, the server determines the product of the initial weight coefficient and the third parameter, and uses this product as the weight coefficient corresponding to the behavior dimension; or, in another optional embodiment, the server determines the sum of the initial weight coefficient and the third parameter, and uses this sum as the weight coefficient corresponding to the behavior dimension. It should be noted that the above-mentioned methods of determining the product as the weight coefficient or the sum as the weight coefficient are exemplary methods of this embodiment. In other embodiments, the server may use other calculation methods that can represent the correlation between the weight coefficient and the initial weight coefficient and the third parameter to determine the weight coefficient. This embodiment does not limit this method. The following embodiments of this disclosure will use the determination of the product as the weight coefficient as an example to illustrate the scheme.

[0151] In some embodiments, for the processes of steps 303 to 305 above, the following optimization algorithm (1) can be used to determine the weight coefficients corresponding to each behavioral dimension. The corresponding process is as follows: For any behavioral dimension, the server determines the weight coefficients corresponding to the behavioral dimension based on the initial weight coefficients corresponding to the behavioral dimension, the first average value of the target object on the behavioral dimension, the second average value of the object group on the behavioral dimension, the second parameter of the target object on the behavioral dimension, and the following optimization algorithm (1).

[0152]

[0153] In the formula, i represents the behavioral dimension i, 0≤i≤K, where K represents the total number of behavioral dimensions, such as 5; u represents the target object; w i This represents the initial weight coefficient corresponding to behavioral dimension i, such as the weight coefficient set by business personnel based on historical experience; The q function represents the optimization function for the weight coefficients, used to output the optimized weight coefficients corresponding to the behavioral dimension i; i,u Let represent the behavior prediction parameters of multiple candidate resources for the target object u on the behavior dimension i, and correspondingly, avg(q) i,u The ) represents the average of the behavior prediction parameters of multiple candidate resources for the target object u in this behavior dimension, which is also the first average; all_user represents the object group; score i,all_user The score represents the behavior prediction parameter of the candidate resources of multiple objects included in the object group on the behavior dimension i, where 0 ≤ score. i,all_user ≤1, correspondingly, avg(score)i,all_user std(q) represents the average of the behavioral prediction parameters of the candidate resources of multiple objects included in the object group on the behavioral dimension i, which is also the second average; i,u ) represents the dispersion of the behavior prediction parameters of multiple candidate resources of the target object u in the behavior dimension i, which is also the second parameter. It should be noted that the fraction shown on the right side of the above optimization algorithm (1) is also the third parameter, and the denominator of the fraction is also the fourth parameter.

[0154] In the above embodiments, based on the positive correlation between the first parameter and the weight coefficient and the positive correlation between the second parameter and the weight coefficient, the third parameter can be quickly determined based on the first parameter and the second parameter. Then, combined with the initial weight coefficient corresponding to the behavior dimension, the weight coefficient corresponding to the behavior dimension can be quickly determined. This improves the efficiency of determining the weight coefficient and also improves the accuracy of determining the weight coefficient.

[0155] For ease of understanding, the optimization algorithm (1) described above will be explained in detail below: The fraction shown on the right side of the optimization algorithm (1) is the optimization factor (or perturbation factor) of the weight coefficient. Multiplying the initial weight coefficient by this optimization factor will yield the optimized weight coefficient. In the numerator of this fraction, it indicates the saliency factor at the object level (i.e., the user level). Specifically, it is the ratio of the first average value corresponding to the target object (i.e., the user who triggered this resource push request) to the second average value corresponding to the object group (i.e., all users). This ratio (i.e., the first parameter) can be used to measure the difference between the target object and the object group. Thus, this ratio reflects the target object's tendency relative to the target group in the corresponding behavioral dimension. For example, if the ratio is greater than 1, it indicates that the target object has a strong tendency (e.g., liking, following) in the corresponding behavioral dimension relative to the average level of the target group, and the weight coefficient of that behavioral dimension can be appropriately increased. If the ratio is less than 1, it indicates that the target object has a weak tendency in the corresponding behavioral dimension relative to the average level of the target group, and the weight coefficient of that behavioral dimension can be appropriately decreased. The denominator of this formula indicates the significance factor at the candidate resource level, specifically the dispersion of the behavioral prediction parameters of multiple candidate resources of the target object in that behavioral dimension, such as the standard deviation. Thus, this dispersion (i.e., the second parameter) can reflect the distribution of multiple candidate resources in this behavioral dimension. For example, taking the standard deviation as an example, if the standard deviation is larger, it means that the distribution of multiple candidate resources in this behavioral dimension is more dispersed. At this time, the candidate resources have more obvious distinguishability and may make a greater contribution to subsequent resource push. The weight coefficient corresponding to this behavioral dimension can be appropriately increased. If the standard deviation is smaller, it means that the distribution of multiple candidate resources in this behavioral dimension is more concentrated. At this time, the candidate resources do not have obvious distinguishability. The weight coefficient corresponding to this behavioral dimension can be appropriately decreased.

[0156] It should be understood that for each resource push request triggered by an object, the push result of that resource push request can be represented by f(u,r), where u represents the current object (e.g., the target object), r represents the resource push request triggered by the current object, and f(u,r) indicates that the push result of the resource push request is related to the current object and the resource push request triggered by the current object. When facing a multi-objective optimization problem for resource push (i.e., with multiple behavioral dimensions), a single request can be represented as a combination of multiple behavioral dimensions. Correspondingly, the push result of this resource push request can be represented as Φ(f(u,q)). i,u )│w i ), where Φ is the fusion function of multiple behavioral dimensions, and q i,u w represents the behavior prediction parameters of the candidate resources of the current object u in the behavior dimension i.i This represents the initial weight coefficient corresponding to behavior dimension i. Therefore, for a single behavior dimension, given a fixed initial weight coefficient, the push result corresponding to that behavior dimension is related to the object u and the candidate resource q. i,u In other words, the differences in the push results of a single request mainly lie in the differences between individual objects and the differences between candidate resources.

[0157] For example, taking likes as an example, if user A and user B each trigger a resource push request, the server, after a recall phase and a coarse-sorting phase, will filter out n candidate videos for user A and user B respectively. Assuming user A likes videos more frequently than user B, the predicted like rate for user A's n candidate videos will generally be higher than the predicted like rate for user B's n candidate videos; this is the difference between individual users. Furthermore, the n candidate videos for the same user will also differ. For example, if user A's n candidate videos all have a like rate of 0.5, while user B's candidate videos have like rates distributed between 0 and 1, we can see that user A's like rate distribution is more concentrated, while user B's like rate distribution is more dispersed; this is also the difference between candidate videos.

[0158] In this embodiment, based on the differences between individual objects and candidate resources mentioned above, a method for optimizing weight coefficients is provided. This method ensures that the weight coefficients for each behavioral dimension are no longer fixed, but rather adaptively adjusted based on the differences between individual objects and candidate resources. This achieves personalized adjustment of the weight coefficients, thereby enabling personalized delivery of multimedia resources. The delivery result of a single request can then be represented as... in That is, the optimization function shown in the above optimization algorithm (1).

[0159] Regarding the process of determining the second parameter, taking the standard deviation as the second parameter as an example, in some embodiments, the server uses the following calculation formula (2) to determine the standard deviation of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension.

[0160]

[0161] In the formula, x represents a random variable, corresponding to the behavior prediction parameters of each candidate resource mentioned in the above embodiments; E(x) represents the average value of the behavior prediction parameters of multiple candidate resources; x 2 It also represents a random variable; E(x) 2 E(x) represents the average of the squared values ​​of the behavioral prediction parameters for multiple candidate resources; it should be noted that E(x)2 )-E(x) 2 This represents the variance of the behavior prediction parameters of multiple candidate resources in the behavior dimension. Thus, through the above calculation formula (2), the mean and standard deviation can be calculated simultaneously in one operation. Therefore, the complexity of the above calculation formula (2) is O(n), where O(n) refers to the complexity of the calculation formula (also called linear time), indicating that the complexity of the calculation formula (2) increases linearly and is proportional to the number of inputs. For the multi-objective optimization problem involving K behavior dimensions involved in the embodiments of this disclosure, the overall complexity of the embodiments of this disclosure can be expressed as O(nK). Considering that the number of candidate resources in a multi-objective optimization problem is usually a constant (n≤500), the complexity of the resource recommendation method provided by this disclosure is O(K). This reduces the computational complexity of the resource recommendation method and improves its computational efficiency.

[0162] In the above embodiments, a method is provided to optimize the initial weight coefficient based on a first parameter and a second parameter to obtain an optimized weight coefficient, so that the determined weight coefficient reflects a positive correlation with both the first and second parameters. It should be noted that the optimization method of the weight coefficient shown above is an exemplary method of this disclosure. In other embodiments, the server may use other calculation methods that can represent the positive correlation between the weight coefficient and the first and second parameters to optimize the initial weight coefficient. For example, the server may determine the product of the initial weight coefficient corresponding to the behavioral dimension, the first parameter, and the second parameter, and determine the product as the optimized weight coefficient. This disclosure does not limit this approach.

[0163] In step 306, the server determines the push parameters of the candidate resource based on the behavior prediction parameters of the candidate resource in the at least two behavior dimensions and the weight coefficients corresponding to the at least two behavior dimensions. The push parameters are used to determine the push order of the candidate resource when pushing resources to the target object.

[0164] The process by which the server determines the push parameters of the candidate resource based on the predicted behavioral parameters of the candidate resource in at least two behavioral dimensions and the corresponding weight coefficients of the at least two behavioral dimensions is described in steps (306A) to (306C):

[0165] (306A) The server sorts the candidate resources according to the behavioral prediction parameters of the candidate resources in the at least two behavioral dimensions.

[0166] In some embodiments, the server sorts the candidate resources in the queues corresponding to the at least two behavioral dimensions according to the behavioral prediction parameters of the at least two behavioral dimensions.

[0167] In one optional embodiment, when the server sorts the behavior prediction parameters for any behavior dimension, if the behavior dimension belongs to the first behavior dimension, it sorts the candidate resources according to the order of the behavior prediction parameters in that behavior dimension from high to low. In other embodiments, if the behavior dimension belongs to the second behavior dimension, it sorts the candidate resources according to the order of the behavior prediction parameters in that behavior dimension from low to high. The first behavior dimension includes positive behavior dimensions and the viewing duration dimension of multimedia resources. Positive behavior dimensions are those corresponding to positive interactive behaviors such as clicking, liking, and saving. The second behavior dimension represents negative behavior dimensions, such as those corresponding to negative interactive behaviors such as blocking and skipping.

[0168] For example, taking at least two behavioral dimensions as the like dimension, the block dimension, and the viewing time dimension of multimedia resources, the corresponding queues for the at least two behavioral dimensions can be the like queue, the block queue, and the viewing time queue of multimedia resources. Then, in the like queue, the candidate resources are sorted in descending order of their like rate; in the block queue, the candidate resources are sorted in ascending order of their block rate; and in the viewing time queue of multimedia resources, the candidate resources are sorted in descending order of their predicted viewing time.

[0169] In the above embodiments, by setting up queues corresponding to each behavioral dimension, the behavioral prediction parameters for each behavioral dimension can be sorted in the queues corresponding to each behavioral dimension.

[0170] (306B) The server determines the push sub-parameters of the candidate resource in the at least two behavioral dimensions based on the order of the candidate resource in the at least two behavioral dimensions.

[0171] The push sub-parameter indicates the ranking of candidate resources in the corresponding behavioral dimension. Taking a queue as an example, the push sub-parameter indicates the ranking of candidate resources within the corresponding queue.

[0172] In this embodiment, the push sub-parameter is negatively correlated with the ranking order. Accordingly, the server determines the push sub-parameter as follows: In one optional embodiment, the server determines the reciprocal of the ranking order of the candidate resource in the at least two behavioral dimensions, and determines the reciprocal as the push sub-parameter of the candidate resource in the at least two behavioral dimensions; or, in another optional embodiment, the server determines a difference between the ranking order of the candidate resource in the at least two behavioral dimensions and the ranking order itself, and determines the difference as the push sub-parameter of the candidate resource in the at least two behavioral dimensions. In this embodiment, by taking the reciprocal or the difference of the ranking order of the candidate resource in each behavioral dimension, the push sub-parameter of the candidate resource in each behavioral dimension can be quickly determined, improving the efficiency of determining the push sub-parameter.

[0173] (306C) The server determines the push parameters of the candidate resource based on the push sub-parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0174] In some embodiments, the server performs a weighted summation based on the push sub-parameters of the candidate resource across at least two behavioral dimensions and the corresponding weight coefficients of those at least two behavioral dimensions to obtain the push parameters for the candidate resource. In this embodiment, the weighted summation method achieves the fusion of behavioral prediction parameters across various behavioral dimensions, enabling the determination of highly accurate push parameters. Resource push is then performed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

[0175] In the above embodiments, the push sub-parameters of the candidate resources in each behavioral dimension are determined based on the order of the candidate resources in each behavioral dimension, so that the determined push sub-parameters can reflect the ranking of the candidate resources. Then, based on the push sub-parameters of the candidate resources in each behavioral dimension and the weight coefficients corresponding to each behavioral dimension, the push parameters of the candidate resources are determined. In this way, by combining the push sub-parameters of the candidate resources in multiple behavioral dimensions, push parameters with high accuracy can be determined, and then resources are pushed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

[0176] In step 307, the server pushes resources to the target object based on the push parameters of the candidate resource.

[0177] In some embodiments, by using steps 301 to 306 above, the push parameters of multiple candidate resources for the target object can be obtained. Then, the multiple candidate resources are sorted in descending order of their push parameters, and the multimedia resources corresponding to the sorted order are pushed to the target object in sequence according to the sorted order of the multiple candidate resources.

[0178] For example, Figure 4 This is a flowchart illustrating a resource push method according to an exemplary embodiment, see [link / reference]. Figure 4 First, the behavior prediction model is used to predict the n candidate videos (i.e., candidate resources) of the target user (i.e., the target object), and the behavior prediction values ​​(i.e., behavior prediction parameters) on K behavior dimensions are obtained respectively. Then, K behavior queues are generated based on the behavior prediction values ​​of the n candidate videos on K behavior dimensions. Then, the above optimization algorithm (1) is used to calculate the perturbation factor corresponding to each behavior dimension, and the initial weight coefficients of each behavior dimension are optimized to obtain the optimized weight coefficients. Specifically, under the condition of confidence of the behavior prediction model, since the overall level of the behavior prediction parameters of the candidate videos on the same behavior dimension can reflect the target user's behavior in the current behavior dimension, the behavior prediction parameters of the target user (i.e., the target object) are predicted by the behavior prediction model. Therefore, the average value of the behavioral prediction of the n candidate videos of the target user is calculated and compared with the average value of the behavioral prediction of multiple users included in the user group (i.e., the target group) to reflect the degree of the target user's tendency in the current behavioral dimension. At the same time, the standard deviation of the behavioral prediction of the n candidate videos of the target user in the current behavioral dimension is calculated to reflect the distribution of multiple candidate videos of the target user in the current behavioral dimension. Then, the initial weight coefficients of each behavioral dimension are optimized using the above optimization algorithm (1) to obtain the optimized weight coefficients. Then, the rank score (i.e., the push sub-parameter) of the candidate video in each behavioral dimension and the optimized weight coefficients of each behavioral dimension are used to determine the ensemble score (i.e., the push parameter) of the candidate video. Based on the above process, the ensemble scores of multiple candidate videos can be obtained. The ensemble scores of multiple candidate videos are sorted, and the resources are pushed according to the order of the multiple candidate resources.

[0179] The technical solution provided in this disclosure, when pushing resources to a target object, determines the weight coefficients corresponding to each behavioral dimension based on the first and second parameters of the target object in each behavioral dimension. Thus, in determining the weight coefficients for each behavioral dimension, not only are the differences of the target object relative to the object group in each behavioral dimension considered, but also the distribution of multiple candidate resources of the target object in each behavioral dimension. This ensures that the determined weight coefficients not only reflect the tendency of the target object in each behavioral dimension, but also the distribution characteristics of the candidate resources in each behavioral dimension. For different objects, the first and second parameters of each object in different behavioral dimensions are referenced to determine the weight coefficients for each object in different behavioral dimensions, enabling the determination of different weight coefficients for different objects. This results in personalized weight coefficients that reflect the object characteristics and resource characteristics of each object, improving both the efficiency and accuracy of weight coefficient determination. Furthermore, based on the behavioral prediction parameters of the target object's candidate resources in each behavioral dimension and the corresponding weight coefficients, the push parameters for the candidate resources are determined, and resources are then pushed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

[0180] It should be noted that, in the fine-tuning stage of resource push, the resource push method provided in this embodiment of the present disclosure improves the usage time and user retention rate of the application, and also increases the effective playback times of multimedia resources and the interaction rate based on multimedia resources.

[0181] Figure 5 This is a block diagram illustrating a resource delivery device according to an exemplary embodiment. See also... Figure 5 The device includes an acquisition unit 501, a weight coefficient determination unit 502, and a push parameter determination unit 503.

[0182] The acquisition unit 501 is configured to acquire behavior prediction parameters of candidate resources of the target object in at least two behavior dimensions, the behavior prediction parameters representing the prediction of the interaction behavior between the target object and the candidate resource in the behavior dimension.

[0183] The weight coefficient determination unit 502 is configured to perform a first parameter and a second parameter on the target object in the behavioral dimension to determine the weight coefficient corresponding to the behavioral dimension. The first parameter represents the difference of the target object relative to the object group in the behavioral dimension, and the second parameter represents the distribution of multiple candidate resources of the target object in the behavioral dimension.

[0184] The push parameter determination unit 503 is configured to execute the push parameters of the candidate resource based on the behavior prediction parameters of the candidate resource in the at least two behavior dimensions and the weight coefficients corresponding to the at least two behavior dimensions. The push parameters are used to determine the push order of the candidate resource when pushing resources to the target object.

[0185] The technical solution provided in this disclosure, when pushing resources to a target object, determines the weight coefficients corresponding to each behavioral dimension based on the first and second parameters of the target object in each behavioral dimension. Thus, in determining the weight coefficients for each behavioral dimension, not only are the differences of the target object relative to the object group in each behavioral dimension considered, but also the distribution of multiple candidate resources of the target object in each behavioral dimension. This ensures that the determined weight coefficients not only reflect the tendency of the target object in each behavioral dimension, but also the distribution characteristics of the candidate resources in each behavioral dimension. For different objects, the first and second parameters of each object in different behavioral dimensions are referenced to determine the weight coefficients for each object in different behavioral dimensions, enabling the determination of different weight coefficients for different objects. This results in personalized weight coefficients that reflect the object characteristics and resource characteristics of each object, improving both the efficiency and accuracy of weight coefficient determination. Furthermore, based on the behavioral prediction parameters of the target object's candidate resources in each behavioral dimension and the corresponding weight coefficients, the push parameters for the candidate resources are determined, and resources are then pushed based on these push parameters, thereby improving the efficiency and accuracy of resource push.

[0186] In some embodiments, the weighting coefficient determining unit 502 includes:

[0187] The first acquisition subunit is configured to acquire the first average value of the target object in the behavioral dimension and the second average value of the object group in the behavioral dimension. The first average value is the average value of the behavioral prediction parameters of the candidate resources of the target object in the behavioral dimension, and the second average value is the average value of the behavioral prediction parameters of the candidate resources of the multiple objects included in the object group in the behavioral dimension.

[0188] The determination subunit is configured to perform the determination of the first parameter based on the first average value and the second average value, wherein the first parameter is positively correlated with the first average value and negatively correlated with the second average value.

[0189] In some embodiments, the weighting coefficient determining unit 502 includes:

[0190] The second acquisition subunit is configured to acquire the dispersion of the behavior prediction parameters of multiple candidate resources of the target object in the behavior dimension, and determine the dispersion as the second parameter.

[0191] In some embodiments, the weighting coefficient determining unit 502 includes:

[0192] The third parameter determination subunit is configured to execute the first and second parameters of the target object in the behavioral dimension to determine the third parameter of the target object in the behavioral dimension, wherein the third parameter is positively correlated with the first parameter and the second parameter.

[0193] The weight coefficient determination subunit is configured to perform the determination of the weight coefficient corresponding to the behavior dimension based on the initial weight coefficient and the third parameter, wherein the weight coefficient is positively correlated with the initial weight coefficient and the third parameter.

[0194] In some embodiments, the third parameter determines the subunit, which is configured to perform:

[0195] Based on the second parameter of the target object in this behavioral dimension, a fourth parameter of the target object in this behavioral dimension is determined, and the fourth parameter is negatively correlated with the second parameter;

[0196] Based on the first and fourth parameters of the target object in this behavioral dimension, the third parameter is determined, and the third parameter is negatively correlated with the fourth parameter.

[0197] In some embodiments, the acquisition unit 501 is configured to perform:

[0198] Based on the object feature information of the target object and the content feature information of the candidate resource, the behavioral prediction parameters of the candidate resource in at least two behavioral dimensions are determined.

[0199] In some embodiments, the acquisition unit 501 is configured to perform:

[0200] The object feature information of the target object and the content feature information of the candidate resource are input into the behavior prediction model corresponding to the behavior dimension. The behavior prediction model predicts the behavior prediction parameters of the candidate resource in the behavior dimension, and the behavior prediction parameters of the candidate resource in the behavior dimension are obtained. The behavior prediction model is trained based on the object feature information of the sample object, the content feature information of the candidate resource of the sample object, and the behavior parameter labels of the candidate resource of the sample object in the behavior dimension.

[0201] In some embodiments, the push parameter determining unit 503 includes:

[0202] The sorting subunit is configured to sort the candidate resource according to the behavior prediction parameters of the candidate resource in the at least two behavioral dimensions, respectively, in the at least two behavioral dimensions.

[0203] The sub-unit is configured to perform a push sub-parameter for the candidate resource in the at least two behavioral dimensions based on the ranking order of the candidate resource in the at least two behavioral dimensions, wherein the push sub-parameter is negatively correlated with the ranking order.

[0204] The determined subunit is further configured to perform push parameters for the candidate resource based on the push sub-parameters of the candidate resource in the at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0205] In some embodiments, the sorting subunit is configured to perform:

[0206] In the queues corresponding to the at least two behavioral dimensions, the candidate resources are sorted according to the behavioral prediction parameters of the at least two behavioral dimensions.

[0207] In some embodiments, the determining subunit is further configured to perform:

[0208] The push parameters of the candidate resource are obtained by weighted summation based on the push sub-parameters of the candidate resource in at least two behavioral dimensions and the weight coefficients corresponding to the at least two behavioral dimensions.

[0209] It should be noted that the resource push device provided in the above embodiments is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the resource push device and the resource push method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0210] The computer device mentioned in the embodiments of this disclosure can be provided as a terminal. Figure 6A structural block diagram of a terminal 600 provided in an exemplary embodiment of this disclosure is shown. The terminal 600 may be a smartphone, tablet computer, MP3 player (Moving Picture Experts Group Audio Layer III), MP4 player (Moving Picture Experts Group Audio Layer IV), laptop computer, or desktop computer. The terminal 600 may also be referred to as a user device, portable terminal, laptop terminal, desktop terminal, or other names.

[0211] Typically, terminal 600 includes a processor 601 and a memory 602.

[0212] Processor 601 may include one or more processing cores, such as a quad-core processor or an octa-core processor. Processor 601 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 601 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 601 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 601 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0213] The memory 602 may include one or more computer-readable storage media, which may be non-transitory. The memory 602 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 602 are used to store at least one program code, which is executed by the processor 601 to implement the terminal execution process in the resource push method provided in the method embodiments of this disclosure.

[0214] In some embodiments, the terminal 600 may also optionally include a peripheral device interface 603 and at least one peripheral device. The processor 601, memory 602, and peripheral device interface 603 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 603 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 604, a display screen 605, a camera assembly 606, an audio circuit 607, a positioning assembly 608, and a power supply 609.

[0215] Peripheral interface 603 can be used to connect at least one I / O (Input / Output) related peripheral device to processor 601 and memory 602. In some embodiments, processor 601, memory 602 and peripheral interface 603 are integrated on the same chip or circuit board; in some other embodiments, any one or two of processor 601, memory 602 and peripheral interface 603 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0216] The radio frequency (RF) circuit 604 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 604 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 604 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 604 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc. The RF circuit 604 can communicate with other terminals through at least one wireless communication protocol. This wireless communication protocol includes, but is not limited to: metropolitan area networks (MANs), various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks (WLANs), and / or WiFi (Wireless Fidelity) networks. In some embodiments, the RF circuit 604 may also include circuitry related to NFC (Near Field Communication), which is not limited in this disclosure.

[0217] Display screen 605 is used to display a UI (User Interface). This UI may include graphics, text, icons, videos, and any combination thereof. When display screen 605 is a touch display screen, it also has the ability to collect touch signals on or above its surface. These touch signals can be input as control signals to processor 601 for processing. In this case, display screen 605 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard. In some embodiments, there may be one display screen 605, disposed on the front panel of terminal 600; in other embodiments, there may be at least two display screens, disposed on different surfaces of terminal 600 or in a folded design; in other embodiments, display screen 605 may be a flexible display screen, disposed on a curved or folded surface of terminal 600. Furthermore, display screen 605 may be configured as a non-rectangular irregular shape, i.e., a non-rectangular screen. Display screen 605 may be made of materials such as LCD (Liquid Crystal Display) or OLED (Organic Light-Emitting Diode).

[0218] The camera assembly 606 is used to acquire images or videos. Optionally, the camera assembly 606 includes a front-facing camera and a rear-facing camera. Typically, the front-facing camera is located on the front panel of the terminal, and the rear-facing camera is located on the back of the terminal. In some embodiments, there are at least two rear-facing cameras, which are any one of a main camera, a depth-sensing camera, a wide-angle camera, and a telephoto camera, to achieve background blurring by fusion of the main camera and the depth-sensing camera, panoramic shooting by fusion of the main camera and the wide-angle camera, VR (Virtual Reality) shooting, or other fusion shooting functions. In some embodiments, the camera assembly 606 may also include a flash. The flash can be a single-color temperature flash or a dual-color temperature flash. A dual-color temperature flash refers to a combination of a warm light flash and a cool light flash, which can be used for light compensation at different color temperatures.

[0219] The audio circuit 607 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, converting the sound waves into electrical signals that are input to the processor 601 for processing, or input to the radio frequency circuit 604 for voice communication. For stereo sound acquisition or noise reduction purposes, multiple microphones may be used, each located at a different part of the terminal 600. The microphone may also be an array microphone or an omnidirectional microphone. The speaker is used to convert the electrical signals from the processor 601 or the radio frequency circuit 604 into sound waves. The speaker may be a conventional diaphragm speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can convert electrical signals not only into audible sound waves but also into inaudible sound waves for purposes such as distance measurement. In some embodiments, the audio circuit 607 may also include a headphone jack.

[0220] The positioning component 608 is used to locate the current geographical location of the terminal 600 in order to enable navigation or LBS (Location Based Service).

[0221] Power supply 609 is used to supply power to the various components in terminal 600. Power supply 609 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery. When power supply 609 includes a rechargeable battery, the rechargeable battery can support wired or wireless charging. The rechargeable battery can also be used to support fast charging technology.

[0222] In some embodiments, the terminal 600 further includes one or more sensors 610. The one or more sensors 610 include, but are not limited to: an accelerometer 611, a gyroscope 612, a pressure sensor 613, a fingerprint sensor 614, an optical sensor 615, and a proximity sensor 616.

[0223] Accelerometer 611 can detect the magnitude of acceleration along the three coordinate axes of a coordinate system established by terminal 600. For example, accelerometer 611 can be used to detect the components of gravitational acceleration along the three coordinate axes. Processor 601 can control display screen 605 to display the user interface in either a landscape or portrait view based on the gravitational acceleration signal acquired by accelerometer 611. Accelerometer 611 can also be used for games or for acquiring user motion data.

[0224] The gyroscope sensor 612 can detect the orientation and rotation angle of the terminal 600. The gyroscope sensor 612, in conjunction with the accelerometer sensor 611, can collect 3D motion data from the user on the terminal 600. Based on the data collected by the gyroscope sensor 612, the processor 601 can perform the following functions: motion sensing (e.g., changing the UI based on the user's tilt), image stabilization during shooting, game control, and inertial navigation.

[0225] The pressure sensor 613 can be disposed on the side bezel of the terminal 600 and / or on the lower layer of the display screen 605. When the pressure sensor 613 is disposed on the side bezel of the terminal 600, it can detect the user's grip signal on the terminal 600, and the processor 601 can perform left / right hand recognition or quick operation based on the grip signal collected by the pressure sensor 613. When the pressure sensor 613 is disposed on the lower layer of the display screen 605, the processor 601 can control the operable controls on the UI interface based on the user's pressure operation on the display screen 605. The operable controls include at least one of button controls, scroll bar controls, icon controls, and menu controls.

[0226] The fingerprint sensor 614 is used to collect the user's fingerprint. The processor 601 identifies the user's identity based on the fingerprint collected by the fingerprint sensor 614, or the fingerprint sensor 614 identifies the user's identity based on the collected fingerprint. When the user's identity is identified as trusted, the processor 601 authorizes the user to perform relevant sensitive operations, including unlocking the screen, viewing encrypted information, downloading software, making payments, and changing settings. The fingerprint sensor 614 can be located on the front, back, or side of the terminal 600. When the terminal 600 has physical buttons or a manufacturer's logo, the fingerprint sensor 614 can be integrated with the physical buttons or manufacturer's logo.

[0227] An optical sensor 615 is used to collect ambient light intensity. In one embodiment, the processor 601 can control the display brightness of the display screen 605 based on the ambient light intensity collected by the optical sensor 615. Specifically, when the ambient light intensity is high, the display brightness of the display screen 605 is increased; when the ambient light intensity is low, the display brightness of the display screen 605 is decreased. In another embodiment, the processor 601 can also dynamically adjust the shooting parameters of the camera assembly 606 based on the ambient light intensity collected by the optical sensor 615.

[0228] The proximity sensor 616, also known as a distance sensor, is typically mounted on the front panel of the terminal 600. The proximity sensor 616 is used to detect the distance between the user and the front of the terminal 600. In one embodiment, when the proximity sensor 616 detects that the distance between the user and the front of the terminal 600 is gradually decreasing, the processor 601 controls the display screen 605 to switch from a screen-on state to a screen-off state; when the proximity sensor 616 detects that the distance between the user and the front of the terminal 600 is gradually increasing, the processor 601 controls the display screen 605 to switch from a screen-off state to a screen-on state.

[0229] Those skilled in the art will understand that Figure 6The structure shown does not constitute a limitation on terminal 600, and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0230] The computer device mentioned in the embodiments of this disclosure can be provided as a server. Figure 7 This is a block diagram of a server according to an exemplary embodiment. The server 700 can vary significantly due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 701 and one or more memories 702. The one or more memories 702 store at least one line of program code, which is loaded and executed by the one or more processors 701 to implement the server-executed process in the resource push method provided in the various method embodiments described above. Of course, the server 700 may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input / output. The server 700 may also include other components for implementing device functions, which will not be elaborated here.

[0231] In an exemplary embodiment, a computer-readable storage medium including program code is also provided, such as a memory 702 including program code, which can be executed by the processor 701 of the server 700 to complete the resource push method described above. Optionally, the computer-readable storage medium may be ROM (Read-Only Memory), RAM (Random Access Memory), CD-ROM (Compact-Disc Read-Only Memory), magnetic tape, floppy disk, and optical data storage device, etc.

[0232] In an exemplary embodiment, a computer program product is also provided, including a computer program that, when executed by a processor, implements the resource push method described above.

[0233] In some embodiments, the computer program involved in this disclosure may be deployed and executed on a computer device, or on multiple computer devices located in one location, or on multiple computer devices distributed in multiple locations and interconnected through a communication network. The multiple computer devices distributed in multiple locations and interconnected through a communication network may constitute a blockchain system.

[0234] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0235] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A resource pushing method, characterized by, The method comprises: obtaining behavior prediction parameters of a candidate resource of a target object in at least two behavior dimensions, the behavior prediction parameters representing a predicted situation of the target object interacting with the candidate resource in the behavior dimensions; determining a weight coefficient corresponding to the behavior dimension based on a first parameter and a second parameter of the target object in the behavior dimension, the first parameter representing a difference of the target object in the behavior dimension relative to a group of objects, and the second parameter representing a distribution of a plurality of candidate resources of the target object in the behavior dimension; sorting the candidate resources in the at least two behavior dimensions according to the behavior prediction parameters of the candidate resources in the at least two behavior dimensions, respectively; determining a push sub-parameter of the candidate resource in the at least two behavior dimensions based on an arrangement order of the candidate resource in the at least two behavior dimensions, the push sub-parameter being negatively correlated with the arrangement order; determining a push parameter of the candidate resource based on the push sub-parameter of the candidate resource in the at least two behavior dimensions and the weight coefficient corresponding to the at least two behavior dimensions, the push parameter being used to determine a push order of the candidate resource when pushing resources for the target object.

2. The resource pushing method of claim 1, wherein, The process of obtaining the first parameter comprises: obtaining a first average value of the target object in the behavior dimension and a second average value of the group of objects in the behavior dimension, the first average value being an average value of the behavior prediction parameters of the candidate resources of the target object in the behavior dimension, and the second average value being an average value of the behavior prediction parameters of the candidate resources of a plurality of objects included in the group of objects in the behavior dimension; determining the first parameter based on the first average value and the second average value, wherein the first parameter is positively correlated with the first average value, and the first parameter is negatively correlated with the second average value.

3. The resource pushing method of claim 1, wherein, The process of obtaining the second parameter comprises: obtaining a dispersion of the behavior prediction parameters of a plurality of candidate resources of the target object in the behavior dimension, and determining the dispersion as the second parameter.

4. The resource pushing method of claim 1, wherein, The process of determining the weight coefficient corresponding to the behavior dimension based on the first parameter and the second parameter of the target object in the behavior dimension comprises: determining a third parameter of the target object in the behavior dimension based on the first parameter and the second parameter of the target object in the behavior dimension, wherein the third parameter is positively correlated with the first parameter, and the third parameter is positively correlated with the second parameter; determining the weight coefficient corresponding to the behavior dimension based on an initial weight coefficient corresponding to the behavior dimension and the third parameter, wherein the weight coefficient is positively correlated with the initial weight coefficient, and the weight coefficient is positively correlated with the third parameter.

5. The resource pushing method of claim 4, wherein, The process of determining the third parameter of the target object in the behavior dimension based on the first parameter and the second parameter of the target object in the behavior dimension comprises: determine a fourth parameter of the target object on the behavior dimension based on the second parameter of the target object on the behavior dimension, the fourth parameter being negatively correlated with the second parameter; determine the third parameter based on the first parameter and the fourth parameter of the target object on the behavior dimension, the third parameter being negatively correlated with the fourth parameter.

6. The resource pushing method of claim 1, wherein, The obtaining the behavior prediction parameter of the candidate resource of the target object on at least two behavior dimensions includes: determining the behavior prediction parameter of the candidate resource on the at least two behavior dimensions based on the object feature information of the target object and the content feature information of the candidate resource.

7. The resource pushing method of claim 6, wherein, The determining the behavior prediction parameter of the candidate resource on the at least two behavior dimensions based on the object feature information of the target object and the content feature information of the candidate resource includes: inputting the object feature information of the target object and the content feature information of the candidate resource into a behavior prediction model corresponding to the behavior dimension, predicting the behavior prediction parameter of the candidate resource on the behavior dimension through the behavior prediction model, and obtaining the behavior prediction parameter of the candidate resource on the behavior dimension, the behavior prediction model being trained based on object feature information of a sample object, content feature information of a candidate resource of the sample object, and a behavior parameter label of the candidate resource of the sample object on the behavior dimension.

8. The resource pushing method of claim 1, wherein, The sorting in the at least two behavior dimensions according to the behavior prediction parameter of the candidate resource on the at least two behavior dimensions respectively includes: sorting in a queue corresponding to the at least two behavior dimensions according to the behavior prediction parameter of the candidate resource on the at least two behavior dimensions respectively.

9. The resource pushing method of claim 1, wherein, The determining the push parameter of the candidate resource based on the push sub-parameter of the candidate resource on the at least two behavior dimensions and the weight coefficient corresponding to the at least two behavior dimensions includes: performing weighted summation based on the push sub-parameter of the candidate resource on the at least two behavior dimensions and the weight coefficient corresponding to the at least two behavior dimensions to obtain the push parameter of the candidate resource.

10. A resource pushing apparatus characterized by comprising: The device includes: an obtaining unit configured to perform obtaining a behavior prediction parameter of a candidate resource of a target object on at least two behavior dimensions, the behavior prediction parameter representing a prediction of the target object interacting with the candidate resource on the behavior dimension; a weight coefficient determination unit configured to determine a weight coefficient corresponding to the behavior dimension based on a first parameter and a second parameter of the target object on the behavior dimension, the first parameter representing a difference of the target object on the behavior dimension relative to an object group, and the second parameter representing a distribution of a plurality of candidate resources of the target object on the behavior dimension; a push parameter determination unit including a sorting sub-unit and a determination sub-unit; the sorting sub-unit is configured to sort in the at least two behavior dimensions according to the behavior prediction parameter of the candidate resource on the at least two behavior dimensions respectively; The determining subunit is configured to determine a push sub-parameter of the candidate resource on each of the at least two behavior dimensions based on an arrangement order of the candidate resource on the at least two behavior dimensions, the push sub-parameter being negatively correlated with the arrangement order; The determining subunit is further configured to determine a push parameter of the candidate resource based on the push sub-parameter of the candidate resource on the at least two behavior dimensions and a weight coefficient corresponding to the at least two behavior dimensions, the push parameter being used to determine a push order of the candidate resource when the resource is pushed to the target object.

11. The resource pushing apparatus according to claim 10, wherein The weight coefficient determining unit comprises: The first obtaining subunit is configured to obtain a first average value of the target object on the behavior dimension and a second average value of the object group on the behavior dimension, the first average value being an average value of a behavior prediction parameter of a candidate resource of the target object on the behavior dimension, and the second average value being an average value of a behavior prediction parameter of a candidate resource of a plurality of objects included in the object group on the behavior dimension; The determining subunit is configured to determine the first parameter based on the first average value and the second average value, wherein the first parameter is positively correlated with the first average value, and the first parameter is negatively correlated with the second average value.

12. The resource pushing apparatus according to claim 10, wherein The weight coefficient determining unit comprises: The second obtaining subunit is configured to obtain a dispersion of behavior prediction parameters of a plurality of candidate resources of the target object on the behavior dimension, and determine the dispersion as the second parameter.

13. The resource pushing apparatus according to claim 10, wherein The weight coefficient determining unit comprises: The third parameter determining subunit is configured to determine a third parameter of the target object on the behavior dimension based on the first parameter and the second parameter of the target object on the behavior dimension, wherein the third parameter is positively correlated with the first parameter, and the third parameter is positively correlated with the second parameter; The weight coefficient determining subunit is configured to determine a weight coefficient corresponding to the behavior dimension based on an initial weight coefficient corresponding to the behavior dimension and the third parameter, wherein the weight coefficient is positively correlated with the initial weight coefficient, and the weight coefficient is positively correlated with the third parameter.

14. The resource pushing apparatus according to claim 13, wherein The third parameter determining subunit is configured to: determine a fourth parameter of the target object on the behavior dimension based on the second parameter of the target object on the behavior dimension, the fourth parameter being negatively correlated with the second parameter; and determine the third parameter based on the first parameter and the fourth parameter of the target object on the behavior dimension, the third parameter being negatively correlated with the fourth parameter.

15. The resource pushing apparatus according to claim 10, wherein The obtaining unit is configured to: determine the behavior prediction parameter of the candidate resource on the at least two behavior dimensions based on the object feature information of the target object and the content feature information of the candidate resource.

16. The resource pushing apparatus according to claim 15, wherein The obtaining unit is configured to: The object feature information of the target object and the content feature information of the candidate resource are input into a behavior prediction model corresponding to the behavior dimension, a behavior prediction parameter of the candidate resource in the behavior dimension is predicted by the behavior prediction model, and the behavior prediction parameter of the candidate resource in the behavior dimension is obtained. The behavior prediction model is trained based on object feature information of a sample object, content feature information of a candidate resource of the sample object, and a behavior parameter label of the candidate resource of the sample object in the behavior dimension.

17. The resource pushing apparatus according to claim 10, wherein The sorting subunit is configured to perform: In the queue corresponding to each of the at least two behavior dimensions, the candidate resources are sorted according to the behavior prediction parameters of the candidate resources in the at least two behavior dimensions.

18. The resource pushing apparatus according to claim 10, wherein The determining subunit is further configured to perform: Based on the push sub-parameters of the candidate resource in the at least two behavior dimensions and the weight coefficients corresponding to the at least two behavior dimensions, the push parameters of the candidate resource are obtained by weighted summation.

19. A computer device, comprising: The computer device comprises: One or more processors; Memory for storing program code executable by the processor; The processor is configured to execute the program code to implement the resource pushing method according to any one of claims 1 to 9.

20. A computer-readable storage medium, characterized in that, When the program code in the computer readable storage medium is executed by the processor of the computer device, the computer device can execute the resource pushing method according to any one of claims 1 to 9.

21. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the resource pushing method according to any one of claims 1 to 9.

Images