Cache resource processing method and device, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides a cache resource processing method and device, electronic equipment, a storage medium and a program product. The method comprises the steps of obtaining account attribute information of a target account and a preset cache attribute set, wherein the cache attribute set comprises multiple pieces of candidate attribute information; performing service attribute prediction on the target account based on the account attribute information and each piece of candidate attribute information to obtain predicted service attribute information corresponding to each piece of candidate attribute information; selecting candidate attribute information matched with the target account from the cache attribute set as target attribute information according to predicted service attribute information corresponding to each piece of candidate attribute information; and deleting the target cache resource from the cache space, wherein the target cache resource is the cache resource of which the preset resource cache attribute meets the target attribute information. According to the embodiment of the invention, by utilizing the technical scheme, the processing modes of cache resources can be enriched.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to a method, apparatus, electronic device, storage medium, and program product for processing cached resources. Background Technology

[0002] Currently, users can watch videos in video applications. To reduce user waiting time, applications typically pre-download and cache a certain number of video resources, and then play the videos based on the cached video resources.

[0003] After caching a video, the cached video resources can be managed based on a pre-set caching duration. For example, the cached video resource can be deleted after the storage duration reaches the pre-set caching duration.

[0004] However, in existing technologies, the processing methods for cached resources are relatively simple and cannot meet the different resource caching needs of different users. Summary of the Invention

[0005] This disclosure provides a method, apparatus, electronic device, storage medium, and program product for processing cached resources, thereby enriching the ways to process cached resources and meeting the personalized processing needs of different users for cached resources.

[0006] In a first aspect, embodiments of this disclosure provide a method for processing cached resources, including:

[0007] Obtain the account attribute information of the target account and a preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of preset resource cache attributes;

[0008] Based on the account attribute information and each of the candidate attribute information, business attribute prediction is performed on the target account to obtain the predicted business attribute information corresponding to each of the candidate attribute information.

[0009] Based on the predicted business attribute information corresponding to each of the candidate attribute information, select candidate attribute information that matches the target account from the cached attribute set, and use the candidate attribute information that matches the target account as the target attribute information.

[0010] If the target cache resource exists in the cache space of the target account, the target cache resource is deleted from the cache space, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

[0011] Secondly, embodiments of this disclosure also provide a processing apparatus for cached resources, comprising:

[0012] The acquisition module is used to acquire the account attribute information of the target account and a preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of preset resource cache attributes;

[0013] The prediction module is used to predict the business attributes of the target account based on the account attribute information and each of the candidate attribute information, respectively, to obtain the predicted business attribute information corresponding to each of the candidate attribute information.

[0014] The information determination module is used to select candidate attribute information that matches the target account from the cached attribute set based on the predicted business attribute information corresponding to each candidate attribute information, and to use the candidate attribute information that matches the target account as the target attribute information.

[0015] The resource deletion module is used to delete the target cache resource from the cache space if the target cache resource exists in the cache space of the target account, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

[0016] Thirdly, embodiments of this disclosure also provide an electronic device, including:

[0017] One or more processors;

[0018] Memory, used to store one or more programs.

[0019] When the one or more programs are executed by the one or more processors, the one or more processors implement the cache resource processing method as described in the embodiments of this disclosure.

[0020] Fourthly, embodiments of this disclosure also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the cache resource processing method as described in embodiments of this disclosure.

[0021] Fifthly, embodiments of this disclosure also provide a computer program product that, when executed by a computer, enables the computer to implement the cache resource processing method as described in embodiments of this disclosure.

[0022] The cache resource processing method, apparatus, electronic device, storage medium, and program product provided in this disclosure use attribute information that matches preset resource cache attributes with the account to process the account's cache resources. This can enrich the cache resource processing methods, improve the matching degree between the attribute information used in the preset resource cache attributes of the account and the account, thereby improving the practicality of the attribute information of the preset resource cache attributes of the account and improving the user experience. Attached Figure Description

[0023] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0024] Figure 1 A flowchart illustrating a method for processing cached resources provided in this embodiment of the disclosure;

[0025] Figure 2 A flowchart illustrating another method for processing cached resources provided in this embodiment of the disclosure;

[0026] Figure 3 This is a schematic diagram illustrating the process of determining target attribute information provided in an embodiment of the present disclosure;

[0027] Figure 4 A framework diagram of a first attribute prediction model provided in an embodiment of this disclosure;

[0028] Figure 5 A structural block diagram of a cache resource processing apparatus provided in an embodiment of this disclosure;

[0029] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0030] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0031] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0032] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0033] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0034] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0035] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0036] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0037] Figure 1 This is a flowchart illustrating a method for processing cached resources according to an embodiment of this disclosure. The method can be executed by a cached resource processing device, which can be implemented in software and / or hardware and can be configured in an electronic device, typically in a computer, mobile phone, or tablet computer; exemplary, it can be configured in a server or terminal device. The cached resource processing method provided in this disclosure is applicable to scenarios where the attribute information of a preset resource cache attribute corresponding to an account is determined based on account attribute information, such as determining the cache attribute of a preset cached resource used by an account to be displayed as the first media resource after a cold start based on account attribute information.

[0038] Taking short video scenarios as an example, short video applications (APPs) are one of the mainstream forms of media used by netizens today, and market competition is becoming increasingly fierce. Short video content recommendation and distribution, as well as the basic user experience, are crucial factors affecting the competitiveness of short video APPs. How to optimize user experience and enhance the market competitiveness of APPs by combining the actual value of specific businesses has become a key focus for video and APP developers.

[0039] Typically, the process of producing and consuming a video involves shooting, compositing, uploading, transcoding, distribution, file selection, and playback. All of these steps collectively affect the basic user experience of the video app.

[0040] In some short video apps, a video playback interface can be displayed and a video can be played by default upon app launch. The waiting time from a cold start of a short video app to the first video playback (i.e., the first video played) is called the first-view time. The length of the first-view time is something users can directly perceive, and optimizing the first-view time can improve the user experience. Currently, a common optimization method is to use cached videos, that is, to cache videos that the user has downloaded but not watched during the user's last use of the app as the first video to be played upon the user's current launch. This caching method can save video service request response time and video download time, ensuring that users can see the first video earlier.

[0041] However, current technologies for handling the caching of first-view videos are relatively simplistic. They typically pre-set a cache expiration time for the first-view video, and then delete it from the cache after that time. However, if the cache expiration time is set too short, frequent deletion of cached videos can drastically increase the probability of cache misses during app launch, making it highly likely that users will have to wait for the first-view video to download after a cold start. Conversely, if the cache expiration time is set too long, the timeliness of the first-view video will be compromised.

[0042] In view of this, the present disclosure provides a method for processing cached resources, which determines the attribute information of a preset resource cache attribute that matches the account based on the account attribute information, and processes the account's cached resources using the attribute information of the preset resource cache attribute that matches the account, so as to enrich the processing methods of cached resources and improve the practicality of the attribute information of the preset resource cache attribute of the account.

[0043] like Figure 1 As shown, the method for processing cached resources provided in this embodiment may include:

[0044] S101. Obtain the account attribute information of the target account and the preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of the preset resource cache attributes.

[0045] The target account can be understood as the account whose attribute information corresponding to the preset resource cache attribute needs to be determined. It can be a registered user account. For example, the cache resource processing method provided in this embodiment can be executed by a server or a target client. When executed by a server, the target account can be a user account that has been registered on the server; when executed by a target client, the target account can be the account logged into by the target client, and the target client can be the client logged into by the target account, such as the application logged into by the target account.

[0046] Account attribute information for a specific account (such as a target account) can be understood as the account's attribute information. This attribute information can be considered a description of the account's characteristics. For example, target account attribute information may include the target account's performance characteristic data, historical business characteristic data, and / or historical display characteristic data of media resources, etc.

[0047] The performance characteristic data can be data describing the performance of media resource display for the target account. For example, this performance characteristic data may include the current performance characteristic data of the target account during this media resource display and / or the historical performance characteristic data of the target account. The current performance characteristic data may include, for example, data such as the duration of buffering, media resource switching time, playback bitrate, playback failure rate, and / or network speed during the media resource display process after the target client's current startup. The media resource switching time can be understood as the time spent switching media resources, such as the time from when the user performs the media resource switching operation to when the first frame of the switched media resource is displayed. The historical performance characteristic data can be understood as the performance characteristic data of the target account before the target client's current startup. For example, the historical performance characteristic data may be the average performance characteristic data of the target account over a past first preset time period, such as the average duration of buffering, the average duration of media resource switching, the average playback bitrate, the average playback failure rate, and / or the average network speed during media resource display over the past first preset time period. The first preset duration can be set as needed, such as 3 days, 7 days and / or 14 days, and the number of first preset durations can be one or more.

[0048] Historical business characteristic data can be understood as the business attribute characteristic data of the target account over the past second preset time period. This business attribute characteristic data can be used to describe the characteristics of the target account's business attribute information, and the business attribute characteristic data may include, but is not limited to, the target account's activity characteristic data. For example, historical business characteristic data may include at least one historical business attribute information of the target account to be integrated, such as the historical average media resource display duration, historical online days, and / or historical average client usage time. The second preset time period can be set as needed, such as 3 days, 7 days, and / or 14 days, and there may be one or more second preset time periods. Taking a second preset time period of 7 days as an example, historical business characteristic data may include the target account's average daily media resource display duration over the past 7 days (i.e., average daily media resource display duration), the number of online days over the past 7 days (i.e., how many days were online in the past 7 days), and / or the average daily runtime over the past 7 days (i.e., average daily online time).

[0049] Historical display characteristic data of media resources can be understood as data describing the display status of media resources of a target account within a third preset time period. Taking the first-time display of media resources as an example, the historical display characteristic data of media resources may include the average startup time of the target client, the average first-time display time of the target account, and / or the average display duration of the first-time display media resources, etc., within the past third preset time period. The average startup time can be understood as the average time spent by the target client to start. The first-time display time can be understood as the average waiting time from the startup of the target client to the display of the first media resource. The first-time display duration of the media resources can be understood as the average display duration of the first media resource after the target client starts. The third preset time period can be set as needed, such as 3 days, 7 days, and / or 14 days, etc., and there can be one or more third preset time periods.

[0050] A cache attribute set can be understood as a set of candidate attribute information for preset resource cache attributes. Preset resource cache attributes can be attributes associated with cached resources and can be used to process cached resources. For example, preset resource cache attributes can be cache condition attributes, cache format attributes, cache location attributes, and / or cache duration attributes, etc. Candidate attribute information for preset resource cache attributes can be selectable attribute information for this preset resource cache attribute. For example, multiple optional attribute values ​​can be pre-set as candidate attribute information for preset resource cache attributes as needed. Different candidate attribute information can correspond to different processing methods for cached resources. For example, different candidate attribute information can correspond to different cache conditions, different cache formats, different cache locations, and / or different cache durations, etc. Optionally, the preset resource cache attribute is a resource cache duration attribute, and different candidate attribute information corresponds to different resource cache durations. For example, the cache attribute set can contain multiple candidate attribute information for resource cache duration attributes. When different candidate attribute information is used, the cached resource can have different resource cache durations. This resource caching duration can be understood as the maximum caching duration of the cached resource. When the actual caching duration of a certain account's cached resource reaches the resource caching duration of that account, the cached resource of that account that has reached the resource caching duration needs to be cleared. That is, the cached resource whose actual caching duration has reached the resource caching duration will be cleared from the cache space of this account.

[0051] In this embodiment, the attribute information of the preset resource cache attribute corresponding to the account can be determined based on the account's attribute information. That is, candidate attribute information that matches the account is selected as the attribute information of the preset resource cache attribute of this account's cached resource based on the account's attribute information.

[0052] Specifically, when it is necessary to determine the attribute information of the preset resource cache attributes corresponding to a certain account (such as a target account), such as when a certain account meets the preset conditions for determining the preset resource cache attributes, the account attribute information of this account can be obtained, such as the performance characteristic data, historical business characteristic data and / or historical display characteristic data of media resources of this account, etc.; and further, the preset cache attribute set can be obtained so that the attribute information of the preset resource cache attributes corresponding to this account can be selected from the cache attribute set based on the attribute information of this account.

[0053] The conditions for determining preset resource cache attributes can be understood as the conditions that trigger the determination of the attribute information of preset resource cache attributes for an account, such as the conditions that trigger the update of the attribute information of preset resource cache attributes for an account. The conditions for determining preset resource cache attributes can be set as needed. For example, the conditions can be set to the arrival of the preset resource cache attribute determination period, the startup of the client logged into by the account, and / or the shutdown of the target client. Therefore, the attribute information of the preset resource cache attributes corresponding to the updated account can be determined periodically, when the client logged into by the account starts, and / or when the client logged into the account closes.

[0054] S102. Based on the account attribute information and each candidate attribute information, perform business attribute prediction on the target account to obtain the predicted business attribute information corresponding to each candidate attribute information.

[0055] In this context, predicted business attribute information can be understood as one or more predicted business attribute information of the target account. Business attribute information describes the business attributes of the target account; these business attributes can be the specific basic characteristics and / or nature of the target account's business or activities. Different business attribute information can correspond to different business attributes of the target account. In an optional example, predicted business attribute information can be business attribute information related to the target account's activity level, such as the target account's activity value or activity gain value, etc.

[0056] In this embodiment, after obtaining the account attribute information of the target account, the business attribute information of the target account can be predicted as the predicted business attribute information corresponding to each candidate resource when each candidate attribute information is used as the attribute information of the preset resource cache attribute corresponding to the target account.

[0057] When using a candidate attribute as the attribute information of a preset resource cache attribute corresponding to the target account, the prediction method for the target account's business attribute information is not limited. For example, a business attribute information prediction model can be pre-trained for predicting business attribute information, and this business attribute information prediction model can be used to predict business attribute information.

[0058] For example, a first business attribute prediction model can be trained in advance, targeting multiple candidate attribute information of preset resource cache attributes in a cache attribute set. This first business attribute prediction model can then be used to predict the business attributes of a target account based on these multiple candidate attribute information; specifically, it can be used to predict the predicted business attribute information corresponding to each of the multiple candidate attribute information. Therefore, when predicting the business attributes of a target account, the account attribute information of the target account can be input into the first business attribute prediction model, and the predicted business attribute information corresponding to each candidate attribute information output by the first business attribute prediction model can be obtained.

[0059] For example, a second business attribute prediction model can be pre-trained for each candidate attribute. Then, these different second business attribute prediction models can be used to predict the business attributes of the target account for this candidate attribute with a preset resource cache attribute. Therefore, when determining the predicted business attribute corresponding to a certain candidate attribute, the account attribute information of the target account can be input into the second business attribute prediction model corresponding to this candidate attribute, and the business attribute information output by this second business attribute prediction model can be obtained as the predicted business attribute corresponding to this candidate attribute.

[0060] For example, a third business attribute prediction model can be pre-trained to predict the business attribute information of a preset resource cache attribute corresponding to an account when using different candidate attribute information. This third business attribute prediction model can then be used to predict the predicted business attribute information corresponding to different candidate attribute information of the preset resource cache attribute (including but not limited to the candidate attribute information currently included in the cache attribute set). Even if the candidate attribute information in the cache attribute set changes, the third business attribute prediction model can still be used to predict the business attributes of the target account based on the changed candidate attribute information. Therefore, when predicting the business attributes of a target account, the account attribute information of the target account and the candidate attribute information of the preset resource cache attribute currently included in the cache attribute set can be input into the third business attribute prediction model, and the predicted business attribute information corresponding to each candidate attribute information output by the third business attribute prediction model can be obtained.

[0061] The training methods for the first business attribute information prediction model, the second business attribute information prediction model, and the third business attribute information prediction model can be flexibly set, and this embodiment does not limit them.

[0062] S103. Based on the predicted business attribute information corresponding to each of the candidate attribute information, select candidate attribute information that matches the target account from the cached attribute set, and use the candidate attribute information that matches the target account as the target attribute information.

[0063] Target attribute information can be understood as the selected candidate attribute information that matches the target account. This target attribute information can be used to process the cached resources of the target account.

[0064] For example, after obtaining the predicted business attribute information corresponding to each candidate attribute information of the preset resource cache attributes in the cache attribute set, the candidate attribute information that matches the target account can be selected from the cache attribute set based on the predicted business attribute information corresponding to each candidate attribute information, and this candidate attribute information can be used as the target attribute information.

[0065] In this embodiment, the method for determining the candidate attribute information that matches the target account can be flexibly set. In an optional example, the candidate attribute information can be determined to match the target account based on whether the predicted business attribute information of the candidate attribute information obtained based on the account attribute information meets the preset conditions. For example, candidate attribute information that meets the preset conditions of the predicted business attribute information can be selected from the cached attribute set as candidate attribute information that matches the target account.

[0066] Optionally, the step of selecting candidate attribute information that matches the target account from the cached attribute set based on the predicted business attribute information corresponding to each candidate attribute information includes: selecting candidate attribute information from the cached attribute set that meets preset conditions for the predicted business attribute information, as candidate attribute information that matches the target account.

[0067] Preset conditions can be understood as conditions used to select candidate attribute information that matches the account, and they can be set as needed. Taking the predicted business attribute information of the target account, including the target account's activity gain value, as an example, the preset conditions could be, for example, that the account's activity gain value is the largest, or that it is among the top c (where c is a positive integer) in the sorting of candidate attribute information according to the account's activity gain value, and so on.

[0068] Taking the predicted business attribute information of the target account, including the activity gain value of the target account, as an example, the account activity gain value corresponding to each candidate attribute information can be predicted based on the account attribute information of the target account. Candidate attribute information whose account activity gain value meets the preset conditions can be obtained from the cached attribute set and used as the target attribute information of the preset resource cache attribute of the target account. For example, the candidate attribute information with the largest account activity gain value can be obtained from the cached attribute set and used as the target attribute information of the preset resource cache attribute of the target account.

[0069] The account activity gain value can be used to describe the change in the target account's activity level after adopting the corresponding candidate attribute information compared to before adopting this candidate attribute information (e.g., compared to when using the original attribute information). For example, the account activity gain value can be the difference between the target account's activity level predicted when adopting the corresponding candidate attribute information and the target account's activity level predicted when using the original attribute information. The original attribute information can be understood as the attribute information of the preset resource caching attribute used by the target account at a certain moment, such as the attribute information used by the target account's preset resource caching attribute before cache resource processing is performed using the cache resource processing method provided in this embodiment; the attribute information used by the target account's preset resource caching attribute before the current determination of the target account's preset resource caching attribute attribute; or, the default attribute information of the target account's preset resource caching attribute set in advance, etc.

[0070] Therefore, in scenarios where the preset resource caching attribute is the resource caching duration attribute, by selecting target attribute information that matches the target account as the attribute information of the target account's preset resource caching attribute, for accounts highly sensitive to the timeliness of media resources, such as accounts with business attributes highly sensitive to timeliness, a shorter resource caching duration can be used to process the account's cached resources, improving the timeliness of the displayed media resources. Conversely, for highly active accounts or inactive accounts that are not sensitive to the timeliness of media resources, a relatively longer resource caching duration can be used to process the account's cached resources, increasing the probability of hitting cached resources after the client starts, and reducing user waiting time. Furthermore, this enriches the ways to set attribute information and improves the practicality of the attribute information used for different accounts.

[0071] S104. If the target cache resource exists in the cache space of the target account, the target cache resource is deleted from the cache space, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

[0072] The target account's cache space can be understood as the space used to store the target account's cached resources. This can be the storage space on the electronic device where the target client is installed, such as the cache space within the electronic device used to store the target account's cached resources. The target account's cached resources can be understood as resources cached by the target account, such as media resources cached by the target account. Optionally, the cached resources at least include preset cached resources, which are used as the first media resource displayed upon the next launch of the target client. The target client is the client logged into by the target account.

[0073] The target client refers to the client logged into by the target account, such as the client executing the cached resource processing method provided in this embodiment, or the client corresponding to the server executing the cached resource processing method provided in this embodiment, etc. The preset cached resource can be a resource cached by the target account and used as the first media resource to be displayed after the target client's next startup, i.e., the target account's first-run media resource. The next startup of the target client can include the next cold start or the next warm start. For example, the next startup of the target client can be the next cold start. In some scenarios, when the target client is closed, the first-run media resource (i.e., the preset cached resource) for the next cold start of the target client can be pre-cached so that this first-run media resource is displayed first after the target client's next cold start, reducing the waiting time for users to watch media resources.

[0074] In this embodiment, after obtaining target attribute information that matches the preset resource cache attributes with the target account, the target attribute information that matches the target account can be used to process the cached resources (including but not limited to preset cached resources) of the target account.

[0075] When processing the cached resources of a target account using target attribute information, specifically, it can be determined whether there are target cached resources in the target account's cache space that satisfy the preset resource cache attributes of this target attribute information. If they exist, the target cached resources that satisfy the preset resource cache attributes of this target attribute information are deleted from the target account's cache space.

[0076] Here, the timing of processing the target cached resources is not limited. For example, it can be determined periodically according to a preset processing cycle whether the target account's cache space contains the target cached resources, and when the existence of the target cached resources is determined, they are deleted from the target account's cache space. This preset processing cycle can be set as needed, for example, it can be set to 1 hour, 5 hours, or 12 hours. When the target attribute information matching the target account is different, this preset processing cycle can be the same or different; this embodiment does not limit this.

[0077] The cache resource processing method provided in this embodiment obtains the account attribute information of the target account and a preset cache attribute set, which contains multiple candidate attribute information of the preset resource cache attributes. Based on the account attribute information of the target account and each candidate attribute information in the cache attribute set, business attribute prediction is performed on the target account to obtain predicted business attribute information corresponding to each candidate attribute information. According to the predicted business attribute information corresponding to each candidate attribute information, candidate attribute information matching the target account is selected from the cache attribute set and used as the target attribute information. If a target cache resource exists in the cache space of the target account whose preset resource cache attributes satisfy the target attribute information, then the target cache resource is deleted from the cache space of the target account. This embodiment utilizes the above technical solution, employing attribute information matching the preset resource cache attributes with the account to process the account's cache resources. This enriches the cache resource processing methods, improves the matching degree between the attribute information used in the preset resource cache attributes and the account, thereby improving the practicality of the attribute information of the preset resource cache attributes and enhancing the user experience.

[0078] Figure 2 This is a flowchart illustrating another method for processing cached resources provided in this embodiment. The solution in this embodiment can be combined with one or more optional solutions in the above embodiments. Optionally, the step of predicting business attributes of the target account based on the account attribute information and each candidate attribute information to obtain predicted business attribute information corresponding to each candidate attribute information includes: for each candidate attribute information, predicting multiple business attributes to be merged for the target account based on the account attribute information and the candidate attribute information to obtain multiple business attribute information to be merged for the target account; performing information fusion processing on the multiple business attribute information to be merged, and using the business attribute information obtained from the information fusion processing as the predicted business attribute information corresponding to the candidate attribute information.

[0079] Correspondingly, such as Figure 2 As shown, the method for processing cached resources provided in this embodiment may include:

[0080] S201. Obtain the account attribute information of the target account and the preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of the preset resource cache attributes.

[0081] S202. For each candidate attribute information, based on the account attribute information and the candidate attribute information, predict multiple business attributes to be integrated for the target account to obtain multiple business attribute information to be integrated for the target account.

[0082] The business attribute to be merged can be understood as the business attribute used to merge and obtain predictive business attribute information. In other words, predictive business attribute information can be obtained by merging the business attribute information of at least one (or more) business attributes to be merged. In some scenarios, the business attribute to be merged may include, for example, the number of days the target account will be online in one or more future time periods (such as the next 3 days, 7 days, or 14 days) and / or the daily runtime within the corresponding time period. The business attribute information to be merged can be understood as the information used to describe the target account for that business attribute to be merged, such as the attribute characteristics or attribute values ​​of the target account for that business attribute to be merged.

[0083] Specifically, after obtaining the account attribute information of the target account and the preset cache attribute set, for each candidate attribute information in the cache attribute set, the system can predict the business attribute information to be integrated for each business attribute to be integrated in the target account when the target account's cache resources are processed using the candidate attribute information as the attribute information of the preset resource cache attribute of the target account. For example, it can predict the number of days the target account will be online and / or the daily runtime in one or more future time periods when the target account's cache resources are processed using the candidate attribute information as the attribute information of the preset resource cache attribute of the target account.

[0084] In this embodiment, when using a candidate attribute information as the attribute information of the preset resource cache attribute of the target account to process the cached resources of the target account, the prediction method of the target account's business attribute information to be merged is not limited. For example, a pre-trained attribute prediction model can be used to predict the target account's business attribute information to be merged.

[0085] In some implementations, the same attribute prediction model can be used to predict the target account's business attribute information to be merged, corresponding to different candidate attribute information. Optionally, the step of predicting multiple business attributes to be merged for the target account based on the account attribute information and the candidate attribute information to obtain multiple business attribute information to be merged for the target account includes: for each business attribute to be merged, obtaining the cached attribute set and the first attribute prediction model corresponding to the business attribute to be merged; inputting the account attribute information into the first attribute prediction model; and using the first attribute prediction model to predict the business attribute information to be merged corresponding to each candidate attribute information in the cached attribute set.

[0086] The first attribute prediction model can be a model used to predict the business attribute information to be merged corresponding to each candidate attribute information in the cached attribute set. In other words, the first attribute prediction model can be used to predict the business attribute information to be merged corresponding to different candidate attribute information in the cached attribute set.

[0087] For example, a first attribute prediction model can be trained in advance for the aforementioned candidate attribute information pre-set in the cached attribute set. Thus, when predicting the business attributes to be merged for a target account, the account attribute information of the target account can be input into the first attribute prediction model, and the business attribute information to be merged corresponding to each candidate attribute information output by the first attribute prediction model can be obtained.

[0088] In the above embodiments, the business attribute information of different business attributes to be merged can be predicted using the same or different first attribute prediction models. Optionally, the business attribute information of different business attributes to be merged can be predicted using different first attribute prediction models to reduce the complexity of the first attribute prediction model and improve the prediction accuracy of the business attribute information to be merged. For example, a first attribute prediction model corresponding to each business attribute to be merged can be pre-trained, and when predicting a certain business attribute to be merged for a target account, the first attribute prediction model corresponding to this business attribute to be merged is obtained, the account attribute information of the target account is input into this first attribute prediction model, and the business attribute information to be merged corresponding to each candidate attribute information in the cached attribute set output by this first attribute prediction model is obtained as the business attribute information to be merged for this business attribute corresponding to each candidate attribute information.

[0089] In other embodiments, different attribute prediction models can be used to predict the target account's business attribute information to be merged, corresponding to different candidate attribute information. Optionally, the step of predicting multiple business attributes to be merged for the target account based on the account attribute information and the candidate attribute information to obtain multiple business attribute information to be merged for the target account includes: for each business attribute to be merged and each candidate attribute information, obtaining a second attribute prediction model corresponding to the candidate attribute information and the business attribute to be merged; inputting the account attribute information into the second attribute prediction model; and using the second attribute prediction model to predict the business attribute information to be merged corresponding to the candidate attribute information, wherein different candidate attribute information and the second attribute prediction model corresponding to the business attribute to be merged are different.

[0090] The second attribute prediction model can be an attribute prediction model corresponding to the candidate attribute information in the cached attribute set. For example, different candidate attribute information in the cached attribute set can correspond to different second attribute prediction models, and the second attribute prediction model corresponding to a certain candidate attribute information can be used to predict the business attribute information to be merged for the target account corresponding to that candidate attribute information.

[0091] For example, a second attribute prediction model corresponding to each candidate attribute in the cached attribute set can be trained in advance. Then, when predicting the business attribute information of a target account, the account attribute information of the target account can be input into the second attribute prediction model corresponding to each candidate attribute, and the business attribute information to be fused corresponding to the candidate attribute information output by each second attribute prediction model can be obtained.

[0092] In the above embodiments, the business attribute information of different business attributes to be merged can be predicted using the same or different second attribute prediction models. Optionally, the business attribute information of different business attributes to be merged can be predicted using different second attribute prediction models to reduce the complexity of the second attribute prediction models and improve the prediction accuracy of the business attribute information to be merged. For example, a second attribute prediction model corresponding to each business attribute to be merged can be pre-trained for each candidate attribute information. When predicting a business attribute to be merged for a target account corresponding to a candidate attribute information, the second attribute prediction model corresponding to this business attribute to be merged for this candidate attribute information is obtained. The account attribute information of the target account is input into this second attribute prediction model, and the business attribute information to be merged output by this second attribute prediction model is obtained as the business attribute information to be merged for this candidate attribute information.

[0093] Furthermore, a third-attribute prediction model can be pre-trained to predict the fusion business attribute information of an account when the cached attribute set contains different candidate attribute information. This third-attribute prediction model can then be used to predict the fusion business attribute information corresponding to different candidate attribute information (including but not limited to the candidate attribute information currently contained in the cached attribute set). Even if the candidate attribute information in the cached attribute set changes, the third-attribute prediction model can still be used to predict the fusion business attribute information of the account based on the changed candidate attribute information. Therefore, when predicting the fusion business attribute information of a target account, the account attribute information of the target account and each candidate attribute information currently contained in the cached attribute set can be input into the third-attribute prediction model, and the fusion business attribute information corresponding to each candidate attribute information output by the third-attribute prediction model can be obtained. The fusion business attribute information for different fusion business attributes can be predicted by the same or different third-attribute prediction models; this embodiment does not limit this.

[0094] In some optional scenarios, taking the prediction of the number of online days and / or the daily runtime of the target account in the next k days (i.e., the business attributes to be integrated) using the first attribute prediction model as an example, a first attribute prediction model A for predicting the number of online days and a first attribute prediction model B for predicting the runtime can be pre-trained. Thus, after obtaining the account attribute information of the target account, this account attribute information can be input into the first attribute prediction model A, and the online days value corresponding to each candidate attribute information output by the first attribute prediction model A can be obtained; and this account attribute information can be input into the first attribute prediction model B, and the runtime value corresponding to each candidate attribute information output by the first attribute prediction model B can be obtained.

[0095] In some implementations, considering the existence of accounts that are online every day (full attendance) and accounts that are not online every day (non-full attendance), to reduce the impact of whether an account is a full attendance account on the prediction results, a preset activation layer with a classification function can be set in the attribute prediction model (such as the first attribute prediction model and / or the second attribute prediction model). This classification function classifies the account based on at least some of its attribute information. For example, this classification function can be used to classify the account based on whether it is a full attendance account and output the probability that the account is a full attendance account, assisting in the prediction of the business attributes to be merged. Optionally, the first attribute prediction model and / or the second attribute prediction model includes a preset activation layer, wherein the activation function of the preset activation layer is a classification function, which is used to classify the target account based on at least some of its account attribute information. This at least some account attribute information can be attribute information used to classify the account; when the classification dimensions are different, the at least some attribute information used may be different. For example, this at least some attribute information may include online frequency information; in this case, the at least some attribute information can be used to classify whether the target account is a full attendance account.

[0096] S203. Perform information fusion processing on the multiple business attribute information to be fused, and use the business attribute information obtained from the information fusion processing as the predicted business attribute information corresponding to the candidate attribute information.

[0097] In this embodiment, for each candidate attribute information in the cached attribute set, after predicting multiple business attribute information to be merged for the target account based on this candidate attribute information, the predicted business attribute information corresponding to this candidate attribute information can be calculated based on these multiple business attribute information to be merged. For example, information fusion processing can be performed on the multiple business attribute information to be merged for the target account predicted based on this candidate attribute information, and the business attribute information obtained from the fusion processing can be used as the predicted business attribute information corresponding to this candidate attribute information. The method for calculating the activity value and / or activity gain value based on the business attribute information to be merged is not limited.

[0098] Taking the prediction of business attribute information including activity gain value as an example, for each candidate attribute information, the activity corresponding to the candidate attribute information can be calculated based on the multiple business attribute information to be merged corresponding to the candidate attribute information, and the activity corresponding to the original attribute information can be calculated based on the above multiple business attribute information to be merged corresponding to the original attribute information; and the difference information (such as difference) between the activity corresponding to the candidate attribute information and the activity corresponding to the original attribute information can be calculated as the activity gain value (i.e., the predicted business attribute information) corresponding to the candidate attribute information.

[0099] In some implementations, the step of performing information fusion processing on the plurality of service attribute information to be fused, and using the service attribute information obtained from the information fusion processing as the predicted service attribute information corresponding to the candidate attribute information, includes: determining the gain information of the service attribute to be fused based on the service attribute information to be fused for each service attribute to be fused corresponding to the candidate attribute information; and performing information fusion processing on the gain information of each service attribute to be fused using a target fusion function to obtain the predicted service attribute information corresponding to the candidate attribute information.

[0100] In the above implementation, the gain information of each service attribute to be merged can be determined first, and then the predicted service attribute information corresponding to the candidate attribute information can be determined based on the gain information of each service attribute to be merged.

[0101] Specifically, for each candidate attribute information and each business attribute to be fused, the gain information of this business attribute can be determined based on the corresponding business attribute information of the candidate attribute information. This process can be repeated to obtain the gain information of multiple business attributes to be fused corresponding to the candidate attribute information. After obtaining the gain information of multiple business attributes to be fused corresponding to the candidate attribute information, a target fusion function can be used to fuse the gain information of at least one of these business attributes to obtain the predicted business attribute information corresponding to the candidate attribute information.

[0102] The target fusion function can be determined based on a Bayesian optimization algorithm, such as at least some of the fusion parameters included in the target fusion function. Optionally, at least some of the fusion parameters in the target fusion function are determined based on a Bayesian optimization algorithm. For example, the functional expression of the target fusion function is shown below:

[0103]

[0104] Where, m j Let f be the j-th candidate attribute information, where 1 ≤ j ≤ M. score (m j ) represents the predicted business attribute information (such as the account activity gain value) corresponding to the j-th candidate attribute information. j (ite i ) represents the gain information of the i-th business attribute to be fused, corresponding to the j-th candidate attribute information. a and b are the hyperparameters of the target fusion function, which can be determined based on the Bayesian optimization algorithm.

[0105] In this embodiment, the target fusion function used by different accounts when determining attribute information can be the same or different, and can be flexibly set according to needs.

[0106] In the above embodiments, the gain information of a certain service attribute to be merged can be understood as the attribute change information of this service attribute to be merged when the target account processes the target account's cached resources using a certain candidate attribute information compared to when the target account processes the target account's cached resources using the original attribute information. For example, the difference between the attribute value of this service attribute to be merged when the target account processes the target account's cached resources using a certain candidate attribute information and the original service attribute value of this service attribute to be merged when the target account processes the target account's cached resources using the original attribute information can be determined as the gain information of this service attribute to be merged corresponding to the candidate attribute information.

[0107] Optionally, determining the gain information of the service attribute to be merged based on the service attribute information to be merged for each service attribute to be merged corresponding to the candidate attribute information includes: calculating the difference information between the service attribute information to be merged and the original service attribute information of the service attribute to be merged for each service attribute to be merged corresponding to the candidate attribute information, and using this difference information as the gain information of the service attribute to be merged.

[0108] Among them, the original business attribute information of a certain business attribute to be integrated can be understood as the business attribute information of the business attribute to be integrated when the cached resources of the target account are processed using the original attribute information. It can be predicted by the attribute prediction model (such as the first attribute prediction model or the second attribute prediction model) corresponding to the business attribute to be integrated.

[0109] S204. Based on the predicted business attribute information corresponding to each of the candidate attribute information, select candidate attribute information that matches the target account from the cached attribute set, and use the candidate attribute information that matches the target account as the target attribute information.

[0110] After obtaining the predicted business attribute information corresponding to each candidate attribute information, candidate attribute information matching the target account can be obtained from the cached attribute set. For example, candidate attribute information that meets the preset conditions can be obtained from the cached attribute set and used as the target attribute information matching the target account.

[0111] S205. If the target cache resource exists in the cache space of the target account, the target cache resource is deleted from the cache space, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

[0112] In this embodiment, after obtaining the target attribute information matching the target account, this target attribute information can be used as the attribute information of the target account's preset resource cache attributes, and the cached resources of the target account can be processed based on this target attribute information. The cached resource processing method provided in this embodiment obtains the predicted business attribute information corresponding to each candidate attribute information by fusing multiple business attribute information to be merged corresponding to each candidate attribute information, and determines the target attribute information of the target account based on the predicted business attribute information corresponding to each candidate attribute information. This can improve the accuracy of the determined target attribute information, thereby further improving the practicality of the determined target attribute information.

[0113] Figure 3 This is a schematic diagram illustrating a process for determining attribute information according to an embodiment of the present disclosure. In some optional embodiments, such as... Figure 3 As shown in the embodiments of this disclosure, the method for processing cached resources may include four parts: data collection, model training, inference prediction, and optimization decision-making. Each part is described in detail below:

[0114] S1, Data Collection.

[0115] S1-1. Determine the specific method for adjusting the cache expiration time (i.e., candidate attribute information).

[0116] For example, multiple different adjustment methods can be pre-set. Without loss of generality, let's assume there are M adjustment methods. For example, different adjustment methods can correspond to different cache expiration times (i.e., resource cache durations). The value of M can be set as needed, such as 3, 4, 5, 6, or 7. The resource cache duration can be set to, for example, 1 day, 3 days, 5 days, 7 days, etc.

[0117] S1-2. Based on the specific method of S1-1, start a small-scale A / B test to obtain the account online status under different cache expiration time adjustments.

[0118] For example, for the control group account in the A / B experiment, a basic cache expiration time (such as original attribute information) is set as the raw data (baseline); and different candidate cache expiration times (i.e., candidate attribute information) are set for different experimental group accounts. The A / B experiment period can be adaptively adjusted according to the situation. Taking a period of 14 days as an example, a small-traffic A / B experiment lasts for 14 days, and the data samples of the first 7 days can be used for model training, and the data samples of the last 7 days can be used for model validation.

[0119] S2, Model Training.

[0120] S2-1. Perform modeling and determine the independent and response variables.

[0121] Taking the processing of first-run media resources as an example, for the specific issue of adjusting the first-run cache validity time, the response variables can be the average number of active days and / or the average daily online time (i.e., runtime) over a past period (such as the past 7 days). The independent variables can be first-run feature data (i.e., historical display feature data of first-run media resources), performance feature data, and / or aggregated values ​​of historical business metrics (i.e., historical business feature data). For example, first-run feature data can be the average of the client startup time, first-run duration, and first-run media resource playback duration over the past 7 days. Performance feature data can include current startup performance feature data, such as the duration of buffering, first frame duration, playback bitrate, playback failure rate, and / or network speed during use after startup; it can also include aggregated values ​​of historical performance metrics (i.e., historical performance feature data), such as the average performance feature data over the past 7 days, the average performance feature data over the past 3 days, etc. Aggregated values ​​of historical business metrics can include the average playback duration over the past 7 days, the activity level over the past 7 days, and / or the average app usage time over the past 7 days, etc.

[0122] S2-2. Based on the independent and response variables determined in S2-1, select an appropriate uplift model and train the model on multiple experimental groups to obtain the uplift model (i.e., the first attribute prediction model).

[0123] It should be noted that each experimental group may need to model multiple response variables. Without loss of generality, assuming there are N response variables, there may be N uplift models. In order to make the subsequent optimization problem less complex, for example, the number of response variables can be no more than 3.

[0124] Figure 4 A framework diagram of a first attribute prediction model provided in an embodiment of this disclosure is shown below. Figure 4 As shown, the gain model (i.e., the first attribute prediction model) can include a shared network (Base Net), a Learning Hidden Unit Contributions Net (LHUC Net), and multiple Tower Nets. The feature vectors output by Base Net and LHUC Net can be multiplied together and then input into each Tower Net respectively.

[0125] The Base Net can include discrete feature modules and numerical feature modules. Each discrete feature module and numerical feature module can be connected to a feature fusion (Concat) layer of the Base Net. After this Concat layer, a batch normalization (BN) layer and at least one fully connected layer (two FC layers are shown as an example in the figure) can be sequentially connected. The discrete feature module can be used to process the input discrete feature data. For example, the discrete feature module can include an input layer, a hash layer, an embedding layer, a fully connected (FC) layer, a dropout layer, and a flatten layer, all connected in sequence. The numerical feature module can be used to process the input numerical feature data. For example, the numerical feature module can include an input layer, a BN layer, a fully connected (FC) layer, a dropout layer, another fully connected (FC) layer, and a dropout layer, all connected in sequence.

[0126] LHUC Net can be used for feature enhancement, such as enhancing feature data that is highly correlated with the response variable. For example, when the response variable is the average number of active days (i.e., average online days), the input to LHUC Net can be the initial refresh feature data and the historical average number of active days; when the response variable is the average online time per day, the input to LHUC Net can be the initial refresh feature data and the historical average app usage time, etc. LHUC Net can include at least one sub-network connected in sequence, and each sub-network can include a sequentially connected FC layer, BN layer, FC layer, and BN layer. Furthermore, the first sub-network can also include an Input layer, the output of which can be connected to the input of the first FC layer in this sub-network.

[0127] Different Tower Nets can be used to perform prediction tasks for different resource caching methods (such as original attribute information and candidate attribute information), and output the predicted activity index values ​​(i.e., the business attribute information to be merged) corresponding to the respective resource caching method. For example, the 0th Tower Net can be used to predict the activity index value of the target account when the cached resources of the target account are processed using the original resource caching method (i.e., original attribute information); the 1st to Nth Tower Nets can be used to predict the activity index values ​​of the target account when the cached resources of the target account are processed using the 1st to Nth candidate resource caching methods (i.e., candidate attribute information).

[0128] For example, a Tower Net may include two sequentially connected fully connected (FC) layers. The output of the second FC layer can be connected to the inputs of three FC layers. The output of the second FC layer can be connected to a cumsum layer, and the output of the third FC layer can be connected to another FC layer. The output of this third FC layer can be connected to a sigmoid layer. The feature vector output by the first FC layer can be added to the feature vector output by the cumsum layer connected to the second FC layer. The resulting feature vector can be multiplied or concatenated with the feature vector output by the sigmoid layer. The resulting feature vector can then be input to the output layer of the Tower Net.

[0129] The model input primarily consists of account feature data, including regular discrete and continuous features. Discrete features are hashed and then embedded, while continuous features are mainly derived from fully connected layers and batch normalization layers. Additionally, an LHUC (Learning Hidden Unit Contributions) layer enhances the features, primarily using historical playback data related to the first viewing, such as first viewing duration, first frame duration, and completion rate of the first media resource. Finally, the outputs of the BaseNet and LHUC networks are multiplied, serving as the input to the TowerNet regression task for each subsequent regression task. Each resource caching method has a TowerNet regression task with a corresponding final output; for example, if the prediction target is the number of active days, the output is the predicted number of active days. Each regression task can consist of two parts: 1) a normal quantile regression task, and 2) predicting whether the current sample is a full-attendance user (i.e., whether the account is classified as a full-attendance user, a 0 / 1 classification, assisting the regression task in prediction). Finally, these two parts are multiplied or concatenated to obtain the final predicted value for the number of active days.

[0130] S3, Reasoning and Prediction.

[0131] For example, the N models obtained in step S2-2 can be used to perform inference prediction on multiple accounts (such as some or all accounts S, etc.) and calculate the uplift value of each account.

[0132] S4, Optimal Decision Making.

[0133] Since there are multiple response variables during the model training phase, such as the number of active days in 7 days, the number of active days in 14 days, the average online time per day in 7 days and / or the average online time per day in 14 days, an optimization decision method can be set to determine the optimal resource caching method.

[0134] S4-1. For the multiple response variable values ​​predicted for each candidate resource caching method, Bayesian optimization is first used to perform multi-objective fusion. The fusion function is as follows:

[0135]

[0136] Where, m j Let f be the j-th candidate resource caching method (i.e., candidate attribute information), where 1 ≤ j ≤ M. score (m j ) represents the account activity gain value (i.e., predicted business attribute information) corresponding to the j-th candidate resource caching method. j (ite i ) represents the gain value (i.e., gain information) of the ith activity metric (i.e., the business attribute to be integrated) corresponding to the j-th candidate resource caching method. For example, assuming there are only two response variables: 7-day active days and 14-day active days, then m j (ite1) and m j (ite2) represents the predicted 7-day and 14-day active days for the j-th candidate resource caching method, respectively. Parameters a and b are parameters that can be determined by Bayesian optimization. This fusion function is the black-box objective function of Bayesian optimization.

[0137] Using the above method, multiple rounds of Bayesian optimization can be performed to finally determine parameters a and b. After determining parameters a and b, the fusion uplift value (i.e., activity gain value) of the target account under each candidate resource caching method can be calculated.

[0138] S4-2. Based on the fused uplift value obtained in S4-1, perform optimal decision-making. When there are no constraints in the decision-making process, the optimal decision can be simply maximized.

[0139]

[0140] Finally, the candidate resource caching method with the largest f is selected as the target resource caching method (i.e., target attribute information) for the target account. For example, the cache expiration time corresponding to this candidate resource caching method is used as the cache expiration time of the first-run media resource for the target account.

[0141] The cache resource processing method provided in this embodiment can be simply summarized into four processes: 1) Start small-scale multi-group experiments to collect training data; 2) Use an efficiency-enhancing modeling approach to learn models for different cache expiration times (i.e., candidate attribute information); 3) Perform inference and prediction based on the model to determine the matching between different cache expiration times and accounts; 4) Establish an optimization model to solve for the optimal cache expiration time (i.e., target attribute information) suitable for the account.

[0142] By collecting behavioral feedback from various accounts when displaying media resources through small-scale experimental data, basic features are constructed for model training. This allows for the capture of each account's sensitivity to the timeliness of first-time browsing and their tolerance for first-time browsing time. An adaptive cache expiration time adjustment method is then developed, thereby improving the basic user experience. Furthermore, this model-predictive intelligent adjustment mode for first-time browsing cache expiration time, compared to a fixed mode, does not require setting rule thresholds, making it more favorable for accounts near those thresholds. In addition, the model can intelligently adjust based on changes in an account's sensitivity to the timeliness of first-time browsing and their tolerance for first-time browsing time. This process can be determined and triggered by feature changes caused by environmental variations. By establishing an optimized model, a smart adjustment scheme for first-time browsing cache expiration time that maximizes global business gains is obtained, while simultaneously improving the user's first-time browsing experience.

[0143] The cache resource processing method provided in this embodiment intelligently configures a suitable cache expiration time for each account by comprehensively considering the different accounts' sensitivity to timeliness and tolerance for initial browsing time, thereby improving the user's basic experience when launching the app. This cache resource processing method is obtained by learning from the autonomous behavioral feedback of accounts, which can determine each account's sensitivity to the timeliness of initial browsing media resources and their tolerance for initial browsing time, thereby improving the user's initial browsing experience when launching the app.

[0144] Figure 5 This is a structural block diagram of a cache resource processing apparatus provided in an embodiment of this disclosure. The apparatus can be implemented by software and / or hardware, and can be configured in an electronic device, typically in a computer, mobile phone, or tablet computer; exemplary, it can be configured in a server or terminal device. It can determine the attribute information of a preset resource cache attribute corresponding to an account based on account attribute information by executing a cache resource processing method. For example, it can determine the cache attribute of a preset cache resource used by the account as the first media resource to be displayed after a cold start based on account attribute information. Figure 5 As shown, the cache resource processing device provided in this embodiment may include: an acquisition module 501, a prediction module 502, an information determination module 503, and a resource deletion module 504, wherein,

[0145] The acquisition module 501 is used to acquire the account attribute information of the target account and a preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of preset resource cache attributes;

[0146] Prediction module 502 is used to predict the business attributes of the target account based on the account attribute information and each of the candidate attribute information, respectively, to obtain the predicted business attribute information corresponding to each of the candidate attribute information.

[0147] The information determination module 503 is used to select candidate attribute information that matches the target account from the cached attribute set according to the predicted business attribute information corresponding to each candidate attribute information, and to use the candidate attribute information that matches the target account as the target attribute information.

[0148] The resource deletion module 504 is used to delete the target cache resource from the cache space when the target cache resource exists in the cache space of the target account, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

[0149] The cache resource processing device provided in this embodiment obtains account attribute information of a target account and a preset cache attribute set through an acquisition module. This cache attribute set contains multiple candidate attribute information for preset resource cache attributes. A prediction module predicts the business attributes of the target account based on both the account attribute information and each candidate attribute information in the cache attribute set, obtaining predicted business attribute information corresponding to each candidate attribute information. An information determination module selects candidate attribute information matching the target account from the cache attribute set based on the predicted business attribute information corresponding to each candidate attribute information, and uses this matching candidate attribute information as the target attribute information. A resource deletion module deletes the target cache resource from the target account's cache space if a target cache resource with preset resource cache attributes that satisfy the target attribute information exists in the target account's cache space. This embodiment utilizes the above technical solution, employing attribute information matching preset resource cache attributes with the account to process the account's cache resources. This enriches the cache resource processing methods, improves the matching degree between the attribute information used in the account's preset resource cache attributes and the account, thereby enhancing the usability of the attribute information in the account's preset resource cache attributes and improving the user experience.

[0150] Optionally, the prediction module 502 includes: an information prediction unit, configured to predict multiple business attributes to be merged for the target account based on the account attribute information and the candidate attribute information for each candidate attribute information, thereby obtaining multiple business attribute information to be merged for the target account; and a fusion processing unit, configured to perform information fusion processing on the multiple business attribute information to be merged, and use the business attribute information obtained from the information fusion processing as the predicted business attribute information corresponding to the candidate attribute information.

[0151] The information prediction unit can be specifically used to: for each business attribute to be merged, obtain a first attribute prediction model corresponding to the cached attribute set and the business attribute to be merged; input the account attribute information into the first attribute prediction model, and use the first attribute prediction model to predict the business attribute information to be merged corresponding to each candidate attribute information in the cached attribute set; or, for each business attribute to be merged and each candidate attribute information, obtain a second attribute prediction model corresponding to the candidate attribute information and the business attribute to be merged; input the account attribute information into the second attribute prediction model, and use the second attribute prediction model to predict the business attribute information to be merged corresponding to the candidate attribute information, wherein different candidate attribute information and the second attribute prediction model corresponding to the business attribute to be merged are different.

[0152] Optionally, the business attribute information of different business attributes to be merged is obtained by using different first attribute prediction models or different second attribute prediction models; and / or, the first attribute prediction model and / or the second attribute prediction model includes a preset activation layer, wherein the activation function of the preset activation layer is a classification function, and the classification function is used to classify the target account based on at least some account attribute information of the target account.

[0153] Optionally, the fusion processing unit includes: a gain determination subunit, configured to determine the gain information of the service attribute to be fused based on the service attribute information to be fused for each service attribute to be fused corresponding to the candidate attribute information; and a fusion processing subunit, configured to perform information fusion processing on the gain information of each service attribute to be fused using a target fusion function to obtain the predicted service attribute information corresponding to the candidate attribute information.

[0154] The gain determination subunit can be specifically used to: for each service attribute to be merged corresponding to the candidate attribute information, calculate the difference information between the service attribute information to be merged and the original service attribute information of the service attribute to be merged, and use it as the gain information of the service attribute to be merged.

[0155] Optionally, at least some of the fusion parameters in the target fusion function are determined based on a Bayesian optimization algorithm.

[0156] Optionally, the information determination module 503 may be specifically used to: select candidate attribute information from the cached attribute set that meets preset conditions for predicting business attribute information, as candidate attribute information that matches the target account.

[0157] Optionally, the cached resources include at least a preset cached resource, which is used to be displayed as the first media resource after the target client is launched again, and the target client is the client logged in by the target account; and / or, the preset resource cache attribute is a resource cache duration attribute, and different candidate attribute information corresponds to different resource cache durations.

[0158] The cache resource processing apparatus provided in this disclosure can execute the cache resource processing method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects for executing the cache resource processing method. Technical details not described in detail in this embodiment can be found in the cache resource processing method provided in any embodiment of this disclosure.

[0159] The following is for reference. Figure 6 The diagram illustrates a structural schematic of an electronic device (e.g., a terminal device or a server) 600 suitable for implementing embodiments of the present disclosure. The terminal device in the embodiments of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0160] like Figure 6 As shown, electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 601, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 602 or a program loaded from storage device 608 into random access memory (RAM) 603. RAM 603 also stores various programs and data required for the operation of electronic device 600. Processing device 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0161] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6An electronic device 600 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0162] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a storage device 608, or installed from a ROM 602. When the computer program is executed by the processing device 601, it performs the functions defined in the methods of embodiments of this disclosure.

[0163] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can 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 of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, 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 device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0164] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0165] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0166] The aforementioned computer-readable medium carries one or more programs. When the aforementioned one or more programs are executed by the electronic device, the electronic device causes the following: it acquires account attribute information of a target account and a preset cache attribute set, the cache attribute set containing multiple candidate attribute information of preset resource cache attributes; it performs business attribute prediction on the target account based on the account attribute information and each of the candidate attribute information, obtaining predicted business attribute information corresponding to each candidate attribute information; it selects candidate attribute information matching the target account from the cache attribute set according to the predicted business attribute information corresponding to each candidate attribute information, and uses the candidate attribute information matching the target account as target attribute information; if a target cache resource exists in the cache space of the target account, it deletes the target cache resource from the cache space, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

[0167] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0168] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0169] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of modules do not, in some cases, constitute a limitation on the unit itself.

[0170] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0171] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer 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 of the foregoing.

[0172] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0173] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0174] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for processing cached resources, characterized in that, include: Obtain the account attribute information of the target account and a preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of preset resource cache attributes; Based on the account attribute information and each of the candidate attribute information, business attribute prediction is performed on the target account to obtain the predicted business attribute information corresponding to each of the candidate attribute information. Based on the predicted business attribute information corresponding to each of the candidate attribute information, select candidate attribute information that matches the target account from the cached attribute set, and use the candidate attribute information that matches the target account as the target attribute information. If the target cache resource exists in the cache space of the target account, the target cache resource is deleted from the cache space, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

2. The method according to claim 1, characterized in that, The step of predicting business attributes of the target account based on the account attribute information and each candidate attribute information, to obtain the predicted business attribute information corresponding to each candidate attribute information, includes: For each candidate attribute information, multiple business attributes to be integrated for the target account are predicted based on the account attribute information and the candidate attribute information to obtain multiple business attribute information to be integrated for the target account. The multiple business attribute information to be merged are subjected to information fusion processing, and the business attribute information obtained from the information fusion processing is used as the predicted business attribute information corresponding to the candidate attribute information.

3. The method according to claim 2, characterized in that, The step of predicting multiple business attributes to be integrated for the target account based on the account attribute information and the candidate attribute information to obtain multiple business attribute information to be integrated for the target account includes: For each service attribute to be merged, obtain the cached attribute set and the first attribute prediction model corresponding to the service attribute to be merged; input the account attribute information into the first attribute prediction model, and use the first attribute prediction model to predict the service attribute information to be merged corresponding to each candidate attribute information in the cached attribute set; or For each business attribute to be integrated and each candidate attribute information, a second attribute prediction model corresponding to the candidate attribute information and the business attribute to be integrated is obtained; the account attribute information is input into the second attribute prediction model, and the second attribute prediction model is used to predict the business attribute information to be integrated corresponding to the candidate attribute information, wherein different candidate attribute information and the second attribute prediction model corresponding to the business attribute to be integrated are different.

4. The method according to claim 3, characterized in that, The attribute information of the services to be merged with different attributes is predicted using different first attribute prediction models or different second attribute prediction models; and / or The first attribute prediction model and / or the second attribute prediction model include a preset activation layer, wherein the activation function of the preset activation layer is a classification function, and the classification function is used to classify the target account based on at least some account attribute information of the target account.

5. The method according to claim 2, characterized in that, The step of performing information fusion processing on the multiple business attribute information to be fused, and using the business attribute information obtained from the information fusion processing as the predicted business attribute information corresponding to the candidate attribute information, includes: Based on the candidate attribute information corresponding to each service attribute to be merged, the gain information of the service attribute to be merged is determined. The gain information of each of the service attributes to be fused is processed by a target fusion function to obtain the predicted service attribute information corresponding to the candidate attribute information.

6. The method according to claim 5, characterized in that, The step of determining the gain information of the service attribute to be merged based on the service attribute information of each service attribute to be merged corresponding to the candidate attribute information includes: For each service attribute to be merged corresponding to the candidate attribute information, the difference information between the service attribute information to be merged and the original service attribute information of the service attribute to be merged is calculated, and used as the gain information of the service attribute to be merged.

7. The method according to claim 5, characterized in that, At least some of the fusion parameters in the target fusion function are determined based on a Bayesian optimization algorithm.

8. The method according to any one of claims 1-7, characterized in that, The step of selecting candidate attribute information matching the target account from the cached attribute set based on the predicted business attribute information corresponding to each of the candidate attribute information includes: Candidate attribute information that meets preset conditions for predicting business attribute information is selected from the cached attribute set and used as candidate attribute information that matches the target account.

9. The method according to any one of claims 1-7, characterized in that, The cached resources include at least a preset cached resource, which is used to be displayed as the first media resource after the target client is launched again. The target client is the client logged in by the target account. and / or The preset resource cache attribute is a resource cache duration attribute, and different candidate attribute information corresponds to different resource cache durations.

10. A processing apparatus for cached resources, characterized in that, include: The acquisition module is used to acquire the account attribute information of the target account and a preset cache attribute set, wherein the cache attribute set contains multiple candidate attribute information of preset resource cache attributes; The prediction module is used to predict the business attributes of the target account based on the account attribute information and each of the candidate attribute information, respectively, to obtain the predicted business attribute information corresponding to each of the candidate attribute information. The information determination module is used to select candidate attribute information that matches the target account from the cached attribute set based on the predicted business attribute information corresponding to each candidate attribute information, and to use the candidate attribute information that matches the target account as the target attribute information. The resource deletion module is used to delete the target cache resource from the cache space if the target cache resource exists in the cache space of the target account, wherein the target cache resource is a cache resource whose preset resource cache attributes satisfy the target attribute information.

11. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the method for processing cache resources according to any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the method for processing cache resources as described in any one of claims 1-9.

13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method for processing cached resources as described in any one of claims 1-9.

Images