Multimedia resource determination method, apparatus, electronic device, and storage medium

Abstract

The application provides a multimedia resource determination method and device, electronic equipment and storage medium, and is applied to the technical field of data processing; in the embodiment of the application, at least one candidate multimedia resource is screened out from a multimedia resource set based on object attribute information of a target object; wherein the multimedia resource set contains at least one multimedia resource with a first threshold value of heat value; the variation parameters of each candidate multimedia resource in the at least one candidate multimedia resource are obtained respectively, wherein the variation parameter is used to represent the difference degree of preference between each historical object for the corresponding multimedia resource; at least one target multimedia resource is screened out from the at least one candidate multimedia resource based on the obtained variation parameter. The accuracy of multimedia resource recommendation is improved, and the determined multimedia resource is more consistent with the preference of the target object.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, apparatus, electronic device, and storage medium for determining multimedia resources. Background Technology

[0002] With the development of internet technology and the increase in the number of multimedia resources, in order to improve the viewing experience of multimedia resources, various multimedia resource applications will recommend potentially interesting multimedia resources to the target audience.

[0003] In the process of recommending multimedia resources, the target multimedia resources to be recommended should be screened first. At present, the selection of target multimedia resources mainly adopts a single screening method, which makes the selected multimedia resources inaccurate.

[0004] In summary, how to accurately select multimedia resources and improve the accuracy of multimedia resource recommendations is a problem that needs to be solved. Summary of the Invention

[0005] This application provides a method, apparatus, electronic device, and storage medium for determining multimedia resources, in order to improve the accuracy of multimedia resource recommendations.

[0006] In a first aspect, embodiments of this application provide a method for determining multimedia resources, the method comprising:

[0007] Based on the object attribute information of the target object, at least one candidate multimedia resource is selected from the multimedia resource set; wherein, the multimedia resource set contains at least one multimedia resource whose popularity value reaches a first threshold.

[0008] The variation parameters of each candidate multimedia resource in at least one candidate multimedia resource are obtained respectively, wherein the variation parameters are used to characterize the degree of difference in preference for the corresponding multimedia resources among the various historical objects.

[0009] Based on the obtained mutation parameters, at least one target multimedia resource is selected from at least one candidate multimedia resource.

[0010] Secondly, embodiments of this application provide a multimedia resource determination apparatus, the apparatus comprising:

[0011] The first filtering unit is used to filter at least one candidate multimedia resource from the multimedia resource set based on the object attribute information of the target object; wherein the multimedia resource set includes at least one multimedia resource whose popularity value reaches a first threshold.

[0012] The obtaining unit is used to obtain the variation parameters of each candidate multimedia resource in at least one candidate multimedia resource, wherein the variation parameters are used to characterize the degree of preference difference between each historical object for the corresponding multimedia resource.

[0013] The second screening unit is used to select at least one target multimedia resource from at least one candidate multimedia resource based on the obtained variation parameters.

[0014] In one possible implementation, the first filtering unit is specifically used for:

[0015] Obtain the matching degree between each multimedia resource in the multimedia resource set and the object attribute information of the target object;

[0016] Based on the obtained matching degree, multimedia resources with a matching degree reaching the second threshold are selected and used as candidate multimedia resources.

[0017] In one possible implementation, the obtaining unit determines the variation parameters corresponding to the candidate multimedia resources in the following manner:

[0018] For each candidate multimedia resource in at least one candidate multimedia resource, obtain the set of historical browsing parameters corresponding to each candidate multimedia resource.

[0019] Based on the acquired sets of historical browsing parameters, the variation parameters corresponding to the candidate multimedia resources are determined; wherein, each set of historical browsing parameters includes at least one of the following: a subset of historical browsing duration and a subset of historical browsing count.

[0020] In one possible implementation, if the historical browsing parameter set includes a subset of historical browsing duration or a subset of historical browsing count, then the unit is specifically used for:

[0021] The variation parameters are determined based on the standard deviation or variance between the various historical browsing parameters in the historical browsing parameter set.

[0022] In one possible implementation, if the historical browsing parameter set includes a subset of historical browsing duration and a subset of historical browsing count, then the unit is specifically used for:

[0023] Based on the standard deviations of the historical browsing duration subset and the historical browsing frequency subset, the corresponding first and second sub-variation parameters are determined respectively; then, the first and second sub-variation parameters are weighted, and the first weighted result is used as the variation parameter; or

[0024] Based on the variances of the subsets of historical browsing duration and the subset of historical browsing frequency, the corresponding first and second sub-variation parameters are determined respectively; and the first and second sub-variation parameters are weighted and the second weighted result is used as the variation parameter.

[0025] In one possible implementation, the second filtering unit is specifically used for:

[0026] For each candidate multimedia resource in at least one candidate multimedia resource, the matching degree of the corresponding multimedia resource is weighted based on the variation parameters corresponding to each candidate multimedia resource to obtain the corresponding matching degree weighted processing result.

[0027] Based on the weighted processing results of each matching degree, candidate multimedia resources whose matching degree weighted processing results reach the third threshold are selected, and the selected candidate multimedia resources are used as target multimedia resources.

[0028] In one possible implementation, after the second filtering unit selects at least one target multimedia resource from at least one set of candidate multimedia resources, it is further configured to:

[0029] Rearrange at least one target multimedia resource;

[0030] Based on the rearranged order, at least one target multimedia resource is recommended to the target object.

[0031] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor, wherein the memory is used to store a computer program; and the processor is used to execute the computer program to implement the multimedia resource determination method provided in embodiments of this application.

[0032] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the multimedia resource determination method provided in embodiments of this application.

[0033] The beneficial effects of this application are as follows:

[0034] This application provides a method, apparatus, electronic device, and storage medium for determining multimedia resources, applied in the field of data processing technology. In this application embodiment, firstly, based on the popularity values ​​of each multimedia resource, at least one multimedia resource with a popularity value reaching a first threshold is selected and formed into a multimedia resource set. Then, based on the object attribute information of the target object, at least one candidate multimedia resource is selected from the multimedia resource set. At this point, candidate multimedia resources that the target object may be potentially interested in are selected. After determining each candidate multimedia resource, a variation parameter corresponding to each candidate multimedia resource is obtained to characterize the degree of difference in historical object preferences. Based on the variation parameter, at least one target multimedia resource is selected from the at least one candidate multimedia resource. A second selection is performed based on the variation parameter to select target multimedia resources with high universality and popularity. By using the implementation method of this application to select multimedia resources, the selected target multimedia resources are more in line with the preferences of the target object and have a higher degree of popularity, further improving the accuracy of multimedia resource recommendations.

[0035] Other features and advantages of this application will be set forth in the following description, and will become apparent from the description or may be learned by practicing the application. The objectives and other advantages of this application can be realized and obtained through the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the architecture of a multimedia resource recommendation system.

[0038] Figure 2 This is a diagram illustrating the display of recommended multimedia resources.

[0039] Figure 3 This is a schematic diagram of an application scenario;

[0040] Figure 4 A flowchart illustrating a method for determining multimedia resources provided in this application embodiment;

[0041] Figure 5 A flowchart illustrating a method for filtering candidate multimedia resources provided in this application embodiment;

[0042] Figure 6A schematic diagram illustrating the determination of a target multimedia resource as provided in an embodiment of this application;

[0043] Figure 7 A schematic diagram illustrating another method for determining target multimedia resources provided in this application embodiment;

[0044] Figure 8 This is a schematic diagram illustrating how to determine the variation parameters corresponding to candidate multimedia resources, as provided in an embodiment of this application.

[0045] Figure 9 A schematic diagram illustrating the determination of multimedia resources based on a recommendation system, as provided in an embodiment of this application;

[0046] Figure 10 A schematic diagram illustrating another method for determining multimedia resources based on a recommendation system, provided as an embodiment of this application;

[0047] Figure 11 A structural diagram of a multimedia resource determination device provided in an embodiment of this application;

[0048] Figure 12 An electronic device provided in an embodiment of this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and beneficial effects of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application 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 application described herein can be implemented in sequences other than those illustrated or described herein.

[0051] The following explanations of some terms used in the embodiments of this application are provided to facilitate understanding by those skilled in the art.

[0052] 1. Multimedia resources: In computer systems, multimedia refers to a human-computer interactive information exchange and dissemination medium that combines two or more media; multimedia resources are the data that needs to be disseminated during human-computer interaction, including but not limited to: text, pictures, photos, sound, animation and video.

[0053] 2. Object attribute information, also known as user attribute information, includes, but is not limited to: gender, age, and language information.

[0054] 3. The variation parameter is used to characterize the degree of difference in preference for multimedia resources among various historical objects.

[0055] The design concept of the embodiments of this application will be briefly introduced below.

[0056] This application's embodiments relate to the process of determining multimedia resources. With the development of internet technology, the number of multimedia resources is constantly increasing, and target users are faced with a massive amount of multimedia resources every day. To improve the viewing experience of multimedia resources for target users, various multimedia resource applications typically filter out a portion of the multimedia resources and then recommend these filtered resources to the target users.

[0057] In related technologies, multimedia resource applications often employ a simple, single-filtering method when selecting multimedia resources. For example, based on the multimedia resource tags associated with the target user's historical browsing history, resources with the same tags that the target user has not clicked are filtered out and recommended to the target user. Alternatively, based on the number of clicks for each multimedia resource, resources with higher click counts that the target user has not clicked are filtered out and recommended to the target user.

[0058] However, the filtering methods provided in related technologies are not accurate enough, and it's impossible to predict whether the target audience will be interested in the filtered multimedia resources. In such cases, multimedia resources that the target audience is not interested in may be recommended, causing confusion and reducing the viewing experience. For example, when filtering multimedia resources based on click counts, there may be cases where multimedia resource creators hire people to inflate play counts or click counts, and there may also be errors in historical data processing, leading to inaccurate multimedia resources filtered based on click counts.

[0059] Clearly, in related technologies, the simple screening of multimedia resources when determining recommended multimedia resources results in low accuracy of multimedia resource recommendations.

[0060] Considering that a preliminary screening can be performed based on the popularity of each multimedia resource in the multimedia resource pool, followed by a second screening based on the object attribute information of the target object, and finally a final screening based on the browsing parameters of historical objects, in order to improve the accuracy of multimedia resource recommendations and make the recommended multimedia resources more in line with the preferences of the target object, this application provides a multimedia resource determination method, apparatus, electronic device, and storage medium.

[0061] In this embodiment, firstly, based on the popularity value of each multimedia resource in the multimedia resource pool, at least one multimedia resource with a popularity value reaching a first threshold is selected from the multimedia resource pool and a multimedia resource set is formed. Then, based on the object attribute information of the target object, at least one candidate multimedia resource is selected from the multimedia resource set. At this time, candidate multimedia resources that the target object may be interested in are selected. After determining each candidate multimedia resource, a variation parameter corresponding to each candidate multimedia resource is obtained to characterize the degree of difference in historical object preferences. Based on the variation parameter, at least one target multimedia resource is selected from the at least one candidate multimedia resource. A second selection is performed based on the variation parameter to select multimedia resources that are liked by most historical objects, that is, to select target multimedia resources with high universality and popularity. The selected target multimedia resources are more in line with the preferences of the target object and have a higher degree of popularity, thereby improving the accuracy of multimedia resource recommendations.

[0062] The method for determining multimedia resources provided in this application can be applied to various multimedia resource recommendation systems, such as video recommendation, article recommendation, news recommendation, browser homepage recommendation, and e-commerce recommendation systems. Please refer to... Figure 1 , Figure 1 This is a schematic diagram of the architecture of a multimedia resource recommendation system. The system includes indexing, recall, coarse ranking, fine ranking, and re-ranking components.

[0063] First, the indexing phase is used to retrieve the various multimedia resources stored in the blocks of the entire blockchain;

[0064] Blockchain is a novel application model of distributed data storage computer technology. Each block in a blockchain can serve as a storage node, which can be a server deployed in different regions, storing multimedia resources. Therefore, multiple servers can share multimedia resources through blockchain, forming a data-sharing system. Furthermore, each server in the data-sharing system has a corresponding node identifier, and each server can store the node identifiers of other servers in the system. This allows for the subsequent broadcasting of generated blocks to other servers in the data-sharing system based on their node identifiers.

[0065] Then, the multimedia resources transmitted from the previous layer are filtered sequentially through recall, coarse sorting, and fine sorting to obtain the filtered target multimedia resources.

[0066] For example, in the recall phase, based on the target audience's interest tags, historical behavior, and click counts of various multimedia resources, the system quickly retrieves multimedia resources that the target audience might be interested in and transmits them to the coarse ranking phase. The coarse ranking phase uses a coarse ranking model to match the content features of the target audience and the multimedia resources, determining the matching degree. Based on this matching degree, the multimedia resources obtained in the recall phase are roughly ranked, further reducing the number of multimedia resources transmitted to the fine ranking phase while maintaining a certain level of accuracy. The fine ranking phase uses comprehensive, multi-dimensional features and a fine ranking model to perform personalized ranking of multimedia resources as accurately as possible, and filters the multimedia resources transmitted in the coarse ranking phase. After obtaining the target multimedia resources, these are transmitted to the re-ranking phase.

[0067] Finally, the selected target multimedia resources are deduplicated, reordered, and their diversity guaranteed through a rearrangement process to prevent similar multimedia resources from appearing consecutively. The rearranged target multimedia resources are then recommended to the target user. At this point, the multimedia resource application installed on the target user's terminal device displays the multimedia resources recommended by the multimedia resource recommendation system. For details, please refer to [link / reference]. Figure 2 , Figure 2 Taking video as an example of multimedia resources, it should be noted that... Figure 2 This is just an example and is not the only way to present it.

[0068] It should be noted that both the coarse-ranking model and the fine-ranking model are trained based on user behavior information, interest information, multimedia resource information, and terminal device usage information. However, the coarse-ranking model needs to quickly select multimedia resources from the multimedia resource pool and transmit them to the fine-ranking stage. The fine-ranking stage then reprocesses the multimedia resources selected in the coarse-ranking stage to ensure accuracy. Therefore, the coarse-ranking model mainly ensures speed, while the fine-ranking model ensures accuracy. Consequently, the number of parameters used in the training of the coarse-ranking model in the coarse-ranking stage is less than the number of parameters used in the training of the fine-ranking model.

[0069] In the embodiments of this application, when determining the target multimedia resources, the recall, coarse ranking and fine ranking processes involve artificial intelligence (AI) and machine learning technologies, and are designed based on speech technology, natural language processing technology and machine learning (ML) in artificial intelligence.

[0070] Artificial intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence.

[0071] Artificial intelligence (AI) studies the design principles and implementation methods of various intelligent machines, enabling them to perceive, reason, and make decisions. AI technology mainly includes computer vision, natural language processing, and machine learning / deep learning. With the research and advancement of AI technology, it is being researched and applied in multiple fields, such as smart homes, intelligent customer service, virtual assistants, smart speakers, intelligent marketing, autonomous driving, robotics, and smart healthcare. It is believed that with further technological development, AI will be applied in even more fields and play an increasingly important role.

[0072] Machine learning is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory, among others. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Compared to data mining, which focuses on finding patterns in large datasets, machine learning emphasizes algorithm design, enabling computers to automatically "learn" patterns from data and use these patterns to predict unknown data.

[0073] Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence. Its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, and inductive learning. Reinforcement learning (RL), also known as reward learning, evaluation learning, or enhancement learning, is one of the paradigms and methodologies of machine learning. It is used to describe and solve problems where an agent learns strategies to maximize rewards or achieve specific goals during its interaction with the environment.

[0074] In this embodiment, corresponding filtering conditions are set in the recall, coarse sorting, and fine sorting processes to filter the multimedia resources transmitted from the previous layer. The filtering conditions may include at least one of the following:

[0075] The popularity value of multimedia resources is used to filter out multimedia resources whose popularity value reaches the first threshold; this allows for the filtering of multimedia resources with high current popularity.

[0076] The browsing parameters of each historical object corresponding to the multimedia resource are used to determine the variation parameters of the corresponding multimedia resource. Based on the browsing parameters of each historical object, the multimedia resource is filtered out. This can distinguish the degree of difference between multimedia resources in historical objects and filter out multimedia resources with better universality.

[0077] The multimedia resource tags of the multimedia resources browsed by the target object are used to filter out the multimedia resources corresponding to the specific multimedia resource tags; this allows filtering out multimedia resources related to the multimedia resources that the target object is interested in.

[0078] Multimedia resources are filtered based on the target object's attribute information, such as region, language, and gender; for example, if the target object's language is Chinese, then multimedia resources in Chinese will be filtered out.

[0079] The operation records of the target object, such as search records and historical records, are used to filter out multimedia resources that match the operation records.

[0080] In this way, multimedia resources that reach the first popularity threshold are initially screened based on the object attribute information of the target object to obtain candidate multimedia resources; then, the candidate multimedia resources are screened again based on the browsing parameters of the historical object to make the selected multimedia resources more in line with the preferences of the target object, improve the accuracy of recommended multimedia resources, and improve the user experience of the target object in using multimedia resources.

[0081] The following is a brief introduction to the application scenarios to which the technical solutions of the embodiments of this application are applicable. It should be noted that the application scenarios described below are only for illustrating the embodiments of this application and are not intended to limit the scope. In specific implementation, the technical solutions provided by the embodiments of this application can be flexibly applied according to actual needs.

[0082] Please refer to Figure 3 , Figure 3 An example application scenario is provided in this application embodiment, which includes a terminal device 10 and a server 20;

[0083] Among them, terminal device 10 is a multimedia resource application and website with multimedia resource playback function installed and running; it can be an electronic device such as a personal computer, mobile phone, tablet computer, laptop, e-book reader, vehicle terminal, etc.

[0084] Server 20 is a server containing the server-side component of the target multimedia player; server 20 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or 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, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0085] In one possible implementation, the terminal device 10 and the server 20 can communicate via a communication network, which can be a wired network or a wireless network. Therefore, the terminal device 10 and the server 20 can be directly or indirectly connected via wired or wireless communication. For example, the terminal device 10 can also be indirectly connected to the server 20 via a wireless access point 12, or the terminal device 10 can be directly connected to the server 20 via the Internet; this application does not impose any limitations on this.

[0086] Based on the above application scenarios, the multimedia resource determination method provided by the exemplary embodiments of this application will be described below with reference to the accompanying drawings. It should be noted that the above application scenarios are only shown to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way in this respect.

[0087] Please refer to Figure 4 , Figure 4 An exemplary embodiment of this application provides a method for determining multimedia resources, applied to a multimedia resource recommendation server. The method includes:

[0088] Step S400: Based on the object attribute information of the target object, at least one candidate multimedia resource is selected from the multimedia resource set; wherein, the multimedia resource set includes at least one multimedia resource whose popularity value reaches a first threshold.

[0089] In the process of recommending multimedia resources to the target audience, the main task is to filter out multimedia resources that the target audience has not clicked on from the large number of multimedia resources contained in the multimedia resource pool, and that are potentially of interest to the target audience.

[0090] In the process of filtering multimedia resources, a commonly used algorithm is the "hotspot algorithm." In personalized multimedia resource recommendation scenarios, hotspot algorithms have a high traffic share for multimedia resources, significantly impacting the viewing experience of the target audience. The hotspot algorithm selects multimedia resources with high popularity values ​​and recommends them to the target audience, while ensuring that the target audience has not clicked on these high-popularity resources.

[0091] When filtering multimedia resources based on popularity algorithms, the popularity score is determined based on dimensions such as time, interaction volume, content, and target audience. Among these, the interaction volume and target audience dimensions have the greatest impact on the popularity score, and these dimensions mainly include parameters such as number of clicks, playback completion rate, and engagement rate.

[0092] Therefore, the popularity value in this application embodiment is mainly determined based on at least one of the following: the number of clicks, playback completion rate, and interaction parameters corresponding to the multimedia resource. In other words, in this application embodiment, multimedia resources can be filtered from the multimedia resource pool based on at least one of the following: the number of clicks, playback completion rate, and interaction parameters corresponding to the multimedia resource; for example, filtering based on the number of clicks means filtering multimedia resources whose number of clicks reaches a first threshold and forming a multimedia resource set.

[0093] Although the multimedia resources in the multimedia resource set are popular, it cannot be guaranteed that the target object will like every multimedia resource in the multimedia resource set. Therefore, in this embodiment of the application, at least one candidate multimedia resource is selected from the multimedia resource set based on the object attribute information of the target object.

[0094] Please refer to Figure 5 , Figure 5 An exemplary flowchart of a method for screening candidate multimedia resources in an implementation of this application is provided, including the following steps:

[0095] Step S500: For each multimedia resource in the multimedia resource set, determine the matching degree between the multimedia resource and the object attribute information;

[0096] The matching degree between multimedia resources and object attribute information is determined by a built-in model in the recommendation server, which is trained based on the known matching degree between multimedia resources and object attribute information.

[0097] Step S501: Select multimedia resources with a matching degree reaching the second threshold as candidate multimedia resources.

[0098] Please refer to Figure 6 This is a schematic diagram illustrating the process of selecting candidate multimedia resources from a multimedia resource pool, as provided in an embodiment of this application.

[0099] In one possible implementation, after determining the matching degree between multimedia resources and object attribute information, the multimedia resources in the multimedia resource set are arranged in descending order of matching degree. Then, multimedia resources with a ranking number greater than a preset ranking number threshold are selected as candidate multimedia resources. Please refer to [reference needed]. Figure 7 , Figure 7An exemplary schematic diagram of another method for filtering candidate multimedia resources is provided in an embodiment of this application.

[0100] It should be noted that the matching degree can only predict the target object's preference for the corresponding multimedia resources. In other words, the target object may not necessarily like the selected candidate multimedia resources.

[0101] For example, in the stage of determining the matching degree between multimedia resources and the object attribute information of the target object, both multimedia resources 1 and 2 have high matching degrees, and the corresponding matching degree values ​​are similar. However, the historical object's preference for multimedia resource 2 is more consistent, while its preference for multimedia resource 1 is polarized. At this point, if both multimedia resource 2 and multimedia resource 1 are recommended and distributed, the probability of multimedia resource 2 being liked will be higher than that of multimedia resource 1. Therefore, if both multimedia resource 1 and multimedia resource 2 are recommended to the target object, there may be a situation where the target object likes multimedia resource 2 but dislikes multimedia resource 1.

[0102] Therefore, after determining the matching degree of the multimedia resources and filtering out candidate multimedia resources based on the matching degree, in order to obtain multimedia resources that the target audience likes more and improve the accuracy of the target multimedia resource recommendation, in this embodiment of the application, corresponding variation parameters are determined for each candidate multimedia resource.

[0103] Step S401: Obtain the variation parameters of each candidate multimedia resource in at least one candidate multimedia resource, wherein the variation parameters are used to characterize the degree of preference difference between each historical object for the corresponding multimedia resource.

[0104] In this embodiment of the application, the variation parameters of the corresponding candidate multimedia resources are determined based on the historical browsing parameter set corresponding to the candidate multimedia resources. The historical browsing parameter set includes at least one of the following: a historical browsing duration subset and a historical browsing count subset. The historical browsing duration subset includes the browsing duration of each historical object, and the historical browsing count subset includes the browsing count of each historical object.

[0105] In one possible implementation, the multimedia resource recommendation server records the browsing logs of each multimedia resource published in the corresponding multimedia resource application. The browsing log for a multimedia resource contains all historical objects that have been clicked on that multimedia resource, as well as at least one of the following: browsing duration and number of views for each historical object.

[0106] It should be noted that the browsing logs for each multimedia resource can be recorded in a table format, as shown in Table 1:

[0107] Table 1

[0108]

[0109] Therefore, before determining the variation parameters corresponding to the multimedia resources, the set of historical browsing parameters corresponding to the multimedia resources should be obtained first.

[0110] Please refer to Figure 8 This will be illustrated using the example of determining the variation parameters corresponding to a candidate multimedia resource. Figure 8 A flowchart of a method for determining the variation parameters corresponding to candidate multimedia resources provided in this application embodiment includes the following steps:

[0111] Step S800: Obtain the set of historical browsing parameters corresponding to the candidate multimedia resources;

[0112] In one possible implementation, the browsing duration of each historical object is obtained from the browsing log corresponding to the multimedia resource, forming a subset of historical browsing durations; and the number of times each historical object is viewed is obtained from the browsing log corresponding to the multimedia resource, forming at least one of the subset of historical browsing counts.

[0113] Step S801: Based on the acquired set of historical browsing parameters, determine the variation parameters corresponding to the candidate multimedia resources.

[0114] In one possible implementation, if only a subset of historical browsing durations or historical browsing counts is obtained, and the variation parameters corresponding to the candidate multimedia resources are determined based on the obtained set of historical browsing parameters:

[0115] First, determine the standard deviation or variance among the various historical browsing parameters included in the historical browsing parameter set;

[0116] Then, determine the variation parameter based on the standard deviation or variance; for example, the variation parameter a = (standard deviation / mean) × 100%, the range of the variation parameter is [-1, 1], or the variation parameter a = standard deviation, or the variation parameter a = variance.

[0117] In one possible implementation, if a subset of historical browsing durations and a subset of historical browsing counts are obtained, and the variation parameters corresponding to the candidate multimedia resources are determined based on the obtained set of historical browsing parameters:

[0118] First, determine the standard deviation of the subset of historical browsing duration and the subset of historical browsing count;

[0119] Then, based on each standard deviation, the corresponding first sub-variation parameter and second sub-variation parameter are determined respectively; for example, the first sub-variation parameter = (standard deviation of browsing time / average) × 100%, and the second sub-variation parameter = (standard deviation of browsing times / average) × 100%.

[0120] Finally, the first sub-mutation parameter and the second sub-mutation parameter are weighted and the first weighted result is used as the mutation parameter.

[0121] Where the weight of the first sub-mutation parameter plus the weight of the second sub-mutation parameter equals 1. Or

[0122] First, determine the variances of the subsets of historical browsing duration and the subsets of historical browsing counts.

[0123] Then, based on each variance, the corresponding first sub-variation parameter and second sub-variation parameter are determined respectively; for example, the first sub-variation parameter = the variance of browsing duration, and the second sub-variation parameter = the variance of browsing times;

[0124] Finally, the first and second sub-mutation parameters are weighted and the resulting weighted result is used as the mutation parameter.

[0125] In one possible implementation, the browsing completion rate, which is determined based on the browsing duration, can also be taken into account when determining the coefficient of variation.

[0126] In one possible implementation, the interaction rate between each candidate multimedia resource and the historical object can also be taken into account when determining the coefficient of variation.

[0127] In recommendation scenarios, the variation parameter is a normalized measure of the dispersion of the probability distribution. It can reflect the degree of difference in the preferences of historical objects for multimedia resources. The smaller the variation parameter, the smaller the difference in preferences for multimedia resources among historical objects, and the more consistent the preferences.

[0128] Step S402: Based on the obtained mutation parameters, at least one target multimedia resource is selected from at least one candidate multimedia resource.

[0129] In one possible implementation, when filtering at least one target multimedia resource from at least one candidate multimedia resource based on the obtained mutation parameters:

[0130] First, based on the obtained coefficient of variation, determine the matching weight corresponding to the candidate multimedia resource; for example, matching weight = e -a, where a is the variation parameter. The larger the variation parameter, the smaller the function value and the smaller the final score, indicating that the differences in preference for candidate multimedia resources among different historical objects are greater, and the stability of the candidate multimedia resources being liked is not high.

[0131] Then, based on the determined matching degree weights, the matching degrees corresponding to the candidate multimedia resources are weighted to obtain the corresponding weighted matching degree processing results; for example, if the matching degree corresponding to candidate multimedia resource 1 is S1, the weighted matching degree processing result obtained is S′1=S1*e -a1 , where a1 is the matching weight corresponding to candidate multimedia resource 1;

[0132] Finally, based on the weighted processing results of each matching degree, candidate multimedia resources whose weighted matching degree processing results reach the third threshold are selected, and these selected candidate multimedia resources are used as target multimedia resources. For details, please refer to [link / reference needed]. Figure 6 .

[0133] In one possible implementation, after determining the matching degree weighted processing result for each candidate multimedia resource, the candidate multimedia resources are sorted based on the matching degree weighted processing result. Then, candidate multimedia resources with a ranking number greater than a ranking number threshold are selected as target multimedia resources. Please refer to [reference needed]. Figure 7 .

[0134] It should be noted that the above method for determining multimedia resources can be applied to the recall, coarse ranking, and fine ranking processes of multimedia resources in recommendation servers. In other words, mutation parameters can be added to the recall, coarse ranking, and fine ranking processes to filter multimedia resources. Please refer to [link / reference]. Figure 9 .

[0135] The three screening processes mentioned above can also be combined during recall, coarse ranking, and fine ranking. For example, during recall, the multimedia resource set can be determined based on the number of clicks; during coarse ranking, candidate multimedia resources can be selected from the multimedia resource set based on the object attribute information of the target object; and during fine ranking, the target multimedia resource can be selected from the candidate multimedia resources based on the mutation parameters. Please refer to [reference needed]. Figure 10 .

[0136] In this embodiment of the application, after selecting at least one target multimedia resource from at least one candidate multimedia resource set, the at least one target multimedia resource is rearranged; and the at least one target multimedia resource is recommended to the target object according to the rearranged order.

[0137] In this embodiment, optimizations to the popularization algorithm and the method for determining target multimedia resources enhance the target audience's liking for recommended multimedia resources, improve the accuracy of recommended multimedia resources, and ensure a better viewing experience for the target audience, resulting in an overall improved experience. Furthermore, this can incentivize creators to produce more similar multimedia resources. Through effective control, the production of universally applicable multimedia resources is guided, ensuring the smooth operation of the multimedia resource production chain.

[0138] Based on the same inventive concept, this application provides a multimedia resource determination device 1100. Figure 11 An exemplary embodiment of this application provides a multimedia resource determination device 1100, which includes:

[0139] The first filtering unit 1101 is used to filter at least one candidate multimedia resource from the multimedia resource set based on the object attribute information of the target object; wherein the multimedia resource set includes at least one multimedia resource whose popularity value reaches a first threshold.

[0140] The obtaining unit 1102 is used to obtain the variation parameters of each candidate multimedia resource in at least one candidate multimedia resource, wherein the variation parameters are used to characterize the degree of preference difference between each historical object for the corresponding multimedia resource.

[0141] The second screening unit 1103 is used to screen at least one target multimedia resource from at least one candidate multimedia resource based on the obtained variation parameters.

[0142] In one possible implementation, the first filtering unit 1101 is specifically used for:

[0143] Obtain the matching degree between each multimedia resource in the multimedia resource set and the object attribute information of the target object;

[0144] Based on the obtained matching degree, multimedia resources with a matching degree reaching the second threshold are selected and used as candidate multimedia resources.

[0145] In one possible implementation, the obtaining unit 1102 determines the variation parameters corresponding to the candidate multimedia resources in the following manner:

[0146] For each candidate multimedia resource in at least one candidate multimedia resource, obtain the set of historical browsing parameters corresponding to each candidate multimedia resource.

[0147] Based on the acquired sets of historical browsing parameters, the variation parameters corresponding to the candidate multimedia resources are determined; wherein, each set of historical browsing parameters includes at least one of the following: a subset of historical browsing duration and a subset of historical browsing count.

[0148] In one possible implementation, if the historical browsing parameter set includes a subset of historical browsing duration or a subset of historical browsing count, then the obtaining unit 1102 is specifically used for:

[0149] The variation parameters are determined based on the standard deviation or variance between the various historical browsing parameters in the historical browsing parameter set.

[0150] In one possible implementation, if the historical browsing parameter set includes a subset of historical browsing duration and a subset of historical browsing count, then the obtaining unit 1102 is specifically used for:

[0151] Based on the standard deviations of the historical browsing duration subset and the historical browsing frequency subset, the corresponding first and second sub-variation parameters are determined respectively; then, the first and second sub-variation parameters are weighted, and the first weighted result is used as the variation parameter; or

[0152] Based on the variances of the subsets of historical browsing duration and the subset of historical browsing frequency, the corresponding first and second sub-variation parameters are determined respectively; and the first and second sub-variation parameters are weighted and the second weighted result is used as the variation parameter.

[0153] In one possible implementation, the second filtering unit 1103 is specifically used for:

[0154] For each candidate multimedia resource in at least one candidate multimedia resource, the matching degree of the corresponding multimedia resource is weighted based on the variation parameters corresponding to each candidate multimedia resource to obtain the corresponding matching degree weighted processing result.

[0155] Based on the weighted processing results of each matching degree, candidate multimedia resources whose matching degree weighted processing results reach the third threshold are selected, and the selected candidate multimedia resources are used as target multimedia resources.

[0156] In one possible implementation, after the second filtering unit 1103 filters at least one target multimedia resource from at least one candidate multimedia resource set, it is further configured to:

[0157] Rearrange at least one target multimedia resource;

[0158] Based on the rearranged order, at least one target multimedia resource is recommended to the target object.

[0159] For ease of description, the above sub-models are divided into units (or modules) according to their functions and described separately. Of course, in implementing this application, the functions of each unit (or module) can be implemented in one or more software or hardware components.

[0160] After introducing the multimedia resource determination method and apparatus of exemplary embodiments of this application, the electronic device of another exemplary embodiment of this application will be introduced next.

[0161] Those skilled in the art will understand that various aspects of this application can be implemented as a system, method, or program product. Therefore, various aspects of this application can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, collectively referred to herein as a "circuit," "module," or "system."

[0162] In one possible implementation, the electronic device provided in this application embodiment may include at least a processor and a memory. The memory stores program code that, when executed by the processor, causes the processor to perform each step of the multimedia resource determination method of the various exemplary embodiments of this application.

[0163] The following reference Figure 12 To describe an electronic device 1200 according to this embodiment of the present application. For example... Figure 12 The electronic device 1200 is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0164] like Figure 12 As shown, the sub-models of the electronic device 1200 may include, but are not limited to: at least each of the processors 1201, at least each of the memories 1202, and a bus 1203 connecting different system sub-models (including memories 1202 and processors 1201).

[0165] Bus 1203 represents one or more of several bus structures, including a memory bus or memory controller, peripheral bus, processor, or local bus using any of the multiple bus structures.

[0166] The memory 1202 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 12021 and / or cache memory 12022, and may further include read-only memory (ROM) 12023.

[0167] The memory 1202 may also include a program / utility 12025 having a set (at least each) of program modules 12024, such program modules 12024 including but not limited to: an operating system, each or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0168] Electronic device 1200 can also communicate with each or more external devices 1204 (e.g., keyboard, pointing device, etc.), and with each or more devices that enable a user to interact with electronic device 1200, and / or with any device that enables electronic device 1200 to communicate with each or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (V / O) interface 1205. Furthermore, electronic device 1200 can also communicate with each or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1206. Figure 12 As shown, network adapter 1206 communicates with other modules for electronic device 1200 via bus 1203. It should be understood that, although... Figure 12 As not shown, other hardware and / or software modules can be used in conjunction with electronic device 1200, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAVD systems, tape drives, and data backup storage systems.

[0169] In some possible implementations, various aspects of the multimedia resource determination method provided in this application can also be implemented in the form of a program product, which includes program code. When the program product is run on an electronic device, the program code is used to cause the electronic device to perform the steps in the multimedia resource determination method according to the various exemplary embodiments of this application described above.

[0170] The program product may employ any combination of one or more readable media. A readable media may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0171] The SMS sending control program product of the embodiments of this application can be a portable compact disc read-only memory (CD-ROM) and include program code, and can run on a computing device.

[0172] A readable signal medium may include a data signal propagated in baseband or as a carrier-submodel, carrying readable program code. This propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting a program for use by or in conjunction with a command execution system, apparatus, or device.

[0173] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0174] Program code for performing the operations of this application can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages.

[0175] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to the embodiments of this application, the feature vectors and functions of the two or more units described above can be embodied in each unit. Conversely, the feature vectors and functions of each unit described above can be further divided and embodied by multiple units.

[0176] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into a single step, and / or each step may be broken down into multiple steps.

[0177] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for determining multimedia resources, characterized in that, The method includes: Based on the object attribute information of the target object, at least one candidate multimedia resource is selected from the multimedia resource set; wherein, the multimedia resource set includes at least one multimedia resource whose popularity value reaches a first threshold. For each candidate multimedia resource, the following steps are performed: obtaining the set of historical browsing parameters corresponding to the candidate multimedia resource, and determining the variation parameters of the candidate multimedia resource based on the standard deviation or variance between each historical browsing parameter in the set of historical browsing parameters; wherein, the variation parameters are used to characterize the degree of preference difference between each historical object for the corresponding multimedia resource; Based on the obtained mutation parameters, at least one target multimedia resource is selected from the at least one candidate multimedia resource.

2. The method as described in claim 1, characterized in that, Based on the object attribute information of the target object, at least one candidate multimedia resource is selected from the multimedia resource set, including: The matching degree between each multimedia resource in the multimedia resource set and the object attribute information of the target object is obtained respectively; Based on the obtained matching degree, multimedia resources with a matching degree reaching the second threshold are filtered out, and the filtered multimedia resources are used as candidate multimedia resources.

3. The method as described in claim 1, characterized in that, The set of historical browsing parameters includes at least one of the following: a subset of historical browsing duration and a subset of historical browsing count.

4. The method as described in claim 3, characterized in that, If the set of historical browsing parameters includes a subset of historical browsing duration and a subset of historical browsing count; The step of determining the variation parameters of the candidate multimedia resources based on the standard deviation among the various historical browsing parameters in the historical browsing parameter set includes: Based on the standard deviations of the respective subsets of historical browsing duration and historical browsing frequency, a first sub-variation parameter and a second sub-variation parameter are determined; then, the first sub-variation parameter and the second sub-variation parameter are weighted, and the first weighted result is used as the variation parameter; or The step of determining the variation parameters of the candidate multimedia resources based on the variance among the various historical browsing parameters in the historical browsing parameter set includes: Based on the variances corresponding to the subsets of historical browsing duration and the subsets of historical browsing counts, a first sub-variation parameter and a second sub-variation parameter are determined respectively; and the first sub-variation parameter and the second sub-variation parameter are weighted, and the obtained second weighted processing result is used as the variation parameter.

5. The method as described in claim 2, characterized in that, The step of selecting at least one target multimedia resource from the at least one candidate multimedia resource based on the obtained mutation parameters includes: For each candidate multimedia resource among the at least one candidate multimedia resources, the matching degree of the corresponding multimedia resource is weighted based on the variation parameters corresponding to each candidate multimedia resource to obtain the corresponding matching degree weighting result. Based on the weighted processing results of each matching degree, candidate multimedia resources whose weighted processing results reach the third threshold are selected, and the selected candidate multimedia resources are used as the target multimedia resources.

6. The method as described in claim 1, characterized in that, After selecting at least one target multimedia resource from the at least one candidate multimedia resource set, the process further includes: The at least one target multimedia resource is rearranged. The at least one target multimedia resource is recommended to the target object according to the rearranged order.

7. A multimedia resource determination device, characterized in that, The device includes: The first filtering unit is used to filter at least one candidate multimedia resource from the multimedia resource set based on the object attribute information of the target object; wherein the multimedia resource set includes at least one multimedia resource whose popularity value reaches a first threshold. The acquisition unit is configured to perform the following for each candidate multimedia resource: acquire the set of historical browsing parameters corresponding to the candidate multimedia resource, and determine the variation parameters of the candidate multimedia resource based on the standard deviation or variance between each historical browsing parameter in the set of historical browsing parameters; wherein, the variation parameters are used to characterize the degree of preference difference between each historical object for the corresponding multimedia resource; The second screening unit is used to select at least one target multimedia resource from the at least one candidate multimedia resource based on the obtained variation parameters.

8. An electronic device, characterized in that, The electronic device includes: a memory and a processor, wherein: The memory is used to store computer programs; The processor is configured to execute the computer program to implement the steps of the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 1 to 6.

Images