Information recommendation method and system, edge-side device and storage medium

Abstract

The present application relates to the technical field of data processing. Disclosed are an information recommendation method and system, an edge-side device and a storage medium. In the information recommendation method provided in the embodiments of the present application, an edge-side device can acquire output effect evaluation information obtained by the edge-side device evaluating an output effect of content of a preset type, and can send the output effect evaluation information to a cloud-side device. The cloud-side device can adjust, on the basis of the output effect evaluation information, a push ratio of the content of the preset type that is pushed to the edge-side device. In this way, content recommendation can be dynamically adjusted on the basis of different edge-side devices to adapt to the edge-side devices, thereby improving the output effect of the edge-side devices and improving user experience.

Description

An information recommendation method, system, end-side device, and storage medium

[0001] This application claims priority to Chinese Patent Application No. 202411749577.2, filed on November 29, 2024, entitled "An Information Recommendation Method, System, End-Side Device and Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of data processing technology, specifically to an information recommendation method, system, end-side device, and storage medium. Background Technology

[0003] Currently, various applications (APPs) on terminal devices (or edge devices) can provide personalized recommendations to users, covering a wide range of formats including articles, music, posts, images, videos, and live streams. The existing recommendation method involves the APP's server generating content (also known as a content recommendation list or ranking list) based on user characteristics (e.g., age, viewing preferences), content popularity, or content value (e.g., commercial value), and sending this content to the edge device, which then displays or plays the content.

[0004] However, when displaying or playing content on the device, there may be issues such as poor display or playback. For example, video content may experience stuttering or failure to play, thus affecting the user experience. Summary of the Invention

[0005] This application provides an information recommendation method, system, end-side device, and storage medium.

[0006] In a first aspect, embodiments of this application provide an information recommendation method for a terminal device. The method includes: obtaining output effect evaluation information of a first content among the content pushed within a first time period; sending the output effect evaluation information to a cloud-side device; and receiving content pushed by the cloud-side device during a second time period. The content push ratio of the first content among the content pushed by the cloud-side device during the second time period is determined by the cloud-side device based on the output effect evaluation information, and the content push ratio is used to characterize the proportion of the first content among the pushed content.

[0007] In some embodiments, the first content may refer to the preset type content mentioned in the embodiments of this application. The preset type content may include one or more types of content, such as video content, music content, e-book content, image content, and any other type of content. In some embodiments, the content types may be further divided. For example, video content may include high frame rate video content, low frame rate video content, etc., which is not limited in the embodiments of this application.

[0008] In some embodiments, the duration of the first time period can be determined according to actual needs. For example, the first time period can be 10 minutes, meaning the edge device can evaluate the output effect every 10 minutes. For instance, the first time period can be the duration corresponding to each rendering of the recommendation page, meaning that an output effect evaluation will be performed every time the recommendation page is rendered.

[0009] The information recommendation method provided in this application allows a terminal device to obtain output effect evaluation information (such as display effect evaluation information, playback effect evaluation information, etc.) obtained by evaluating the output effect of preset content types on the terminal device, and then send the output effect evaluation information to a cloud device. The cloud device can adjust the proportion of preset content types pushed to the terminal device based on the output effect evaluation information. For example, if the playback effect evaluation information of video content sent by a terminal device to the cloud device over a certain period indicates that the video content plays poorly on that terminal device, the cloud device can reduce the proportion of video content pushed to that terminal device in subsequent updates. In this way, content recommendations can be dynamically adjusted based on different terminal devices, adapting to the terminal devices, improving the output effect of the terminal devices, and enhancing the user experience.

[0010] In one possible implementation, the first content includes multiple types of content, and the output effect evaluation information of the first content includes output effect scores corresponding to each of the multiple types of content; the cloud-side device determines the content push ratio of the first content in the second time period based on the output effect evaluation information, including: the cloud-side device determines the content push ratio corresponding to each of the multiple types of content in the second time period based on the output effect scores corresponding to each of the multiple types of content.

[0011] In some embodiments, the output performance evaluation information for preset content types may include output performance scores for preset content types within a preset time period. For example, a higher output performance score indicates a better output performance, while a lower score indicates a worse output performance. Alternatively, it can be set so that a lower score indicates a better output performance, and a higher score indicates a worse output performance. When the cloud device subsequently sends content to the endpoint device, it can determine the push ratio of each content type based on the output performance scores. In this way, content recommendations can be dynamically adjusted based on different endpoint devices, improving the output performance of the endpoint devices and enhancing the user experience.

[0012] In one possible implementation, obtaining the output effect evaluation information of the first content among the content pushed within the first time period includes: obtaining the scores of each feature parameter of the first content among the content pushed within the first time period and the weights corresponding to each feature parameter; and determining the output effect score of the first content based on the scores of each feature parameter of the first content and the weights corresponding to each feature parameter.

[0013] In some embodiments, the output performance score of the preset content type can be determined based on the scores of each feature parameter of the preset content type and the weights corresponding to each feature parameter. For example, for the preset content type being video content, the feature parameters of the video content type include at least one of the following: first frame playback status, playability status, frame rate status, stuttering status, central processing unit (CPU) usage status, and memory usage status.

[0014] In some embodiments, the weight of each feature parameter can be determined based on its impact on the output effect of each type of content (i.e., its impact on the user experience). The greater the impact, the greater the weight. For example, for video content of the preset type, excessive delay in the first frame or inability to play the video content has a significant impact on the user experience. That is, the first frame playback status and playability have a significant impact on the output effect of the video content, so a larger weight can be assigned to the first frame playback status and playability status. In this way, a more accurate output effect score can be obtained, thereby enabling precise adjustment of content recommendations, improving the adaptability of recommended content to different terminal devices, and improving the user experience.

[0015] In one possible implementation, the first content includes video content, and the characteristic parameters of the video content include at least one of the following: first frame start-up status, playability status, frame rate status, stuttering status, CPU usage status, and memory usage status.

[0016] In one possible implementation, obtaining scores for each feature parameter of the first content pushed within a first time period includes at least one of the following: determining a score for the first frame playback status of the first video content based on the difference between the first frame display time corresponding to the first video content within the first time period and the time when the terminal device detects the first operation triggering the playback of the first video content; obtaining a score for the playability of the first video content based on whether the first video content was successfully played within the first time period; determining a score for the frame rate of the first video content based on the difference between the frame rate of the first video content and the refresh rate of the terminal device's display screen within the first time period; obtaining a score for the stuttering status of the first video content based on the number of stutters in the first time period; determining a score for the CPU usage of the first video content based on the ratio of the CPU resources occupied by the first video content during playback to the total CPU resources of the terminal device within the first time period; and determining a score for the memory usage of the first video content based on the ratio of the memory resources occupied by the first video content during playback to the total memory resources of the terminal device within the first time period.

[0017] In some embodiments, when the video type content within the first time period includes a video content, for example, including a first video content, the scores of each feature parameter of that video content can be used as the scores of each feature parameter of the video type content within the first time period.

[0018] In some embodiments, when the video content within the first time period includes multiple video contents, the score of each feature parameter of the video content within the first time period can be determined based on the score of each feature parameter of the multiple video contents. For example, the average score of each feature parameter of the multiple video contents can be used as the score of each feature parameter of the video content within the first time period, or the highest and lowest scores among the scores corresponding to the multiple video contents can be removed, and the average score of each feature parameter of the remaining video contents can be used as the score of each feature parameter of the video content within the first time period, etc. The embodiments of this application are not limited to this.

[0019] It is understandable that, based on the first time period, the scores of each feature parameter of multiple video content within the first time period can be used to determine the video type content within the first time period. This can achieve a more accurate output effect score, thereby enabling precise adjustment of content recommendations, improving the adaptability of recommended content to different end devices, and enhancing the user experience.

[0020] In one possible implementation, obtaining the scores of each feature parameter of the first content among the content pushed within the first time period includes: obtaining the scores of each feature parameter corresponding to multiple video content within the first time period; and determining the scores of each feature parameter of the video type content within the first time period based on the scores of each feature parameter corresponding to multiple video content within the first time period.

[0021] In one possible implementation, the first content includes audio content, and the characteristic parameters of the audio content include at least one of the following: first frame start-up status, playability status, stuttering status, CPU usage status, and memory usage status.

[0022] In one possible implementation, the first content includes e-book type content, and the characteristic parameters of the e-book type content include at least one of the following: homepage display status, opening status, lag status, CPU usage, and memory usage.

[0023] In one possible implementation, the first content includes image-type content, and the characteristic parameters of the image-type content include at least one of image loading time, opening status, lag status, CPU usage, and memory usage.

[0024] In one possible implementation, the output effect evaluation information of the first content includes the scores of each feature parameter corresponding to the first content; the method by which the cloud-side device determines the content push ratio of the first content in the second time period based on the output effect evaluation information includes: the cloud-side device obtaining the weights of each feature parameter corresponding to the first content in the first time period; the cloud-side device determining the output effect score of the first content based on the scores of each feature parameter corresponding to the first content and the weights of each feature parameter corresponding to the first content; and the cloud-side device determining the content push ratio of the first content in the second time period based on the output effect score of the first content.

[0025] In some embodiments, the output effect evaluation information corresponding to the preset type of content within a first time period sent by the edge device may include the scores of each feature parameter corresponding to the preset type of content within the first time period. In this embodiment, the cloud device can determine the output effect score corresponding to the preset type of content based on the scores and weights of each feature parameter corresponding to the preset type of content, and determine the content push ratio within a second time period based on the output effect score corresponding to the preset type of content. This reduces the computational load and power consumption of the edge device.

[0026] In one possible implementation, the method further includes: when the first application of the terminal device starts, the terminal device sends a first request to the cloud device corresponding to the first application. The first request is used to obtain evaluation strategy information, which includes each feature parameter of the first content and the weight corresponding to each feature parameter; the cloud device sends the evaluation strategy information to the terminal device; wherein, obtaining the output effect evaluation information of the first content in the content pushed within the first time period includes: obtaining the output effect evaluation information of the first content in the content pushed within the first time period based on the evaluation strategy information.

[0027] In some embodiments, the edge device can send a first request to the cloud-side device corresponding to the information recommendation application (e.g., a first application) when it detects that the application has been launched. After receiving the evaluation strategy information sent by the cloud-side device, the edge device obtains the output effect evaluation information of preset content types within a first time period based on the evaluation strategy information. Thus, the edge device does not need to have the evaluation strategy information built in, saving memory on the edge electronic device.

[0028] In some embodiments, the evaluation strategy information may include feature parameters of preset content types, the weights corresponding to each feature parameter, the evaluation interval duration (e.g., the duration corresponding to the first time period), the evaluation effective time, and the specific calculation method for outputting the evaluation information. The information recommendation application can refer to any application with information recommendation functionality. The cloud-side device may include the server corresponding to the information recommendation application.

[0029] In one possible implementation, obtaining the output effect evaluation information of the first content among the content pushed within the first time period includes: obtaining the output effect evaluation information of the first content among the content pushed within the first time period based on the evaluation strategy information stored in the terminal device, wherein the evaluation strategy information includes each feature parameter of the first content and the weight corresponding to each feature parameter.

[0030] In some embodiments, evaluation strategy information can also be pre-stored in the end-side device, so that the end-side device does not need to obtain it from the cloud-side device, thus saving communication resources.

[0031] In one possible implementation, the cloud-side device determines the content push ratio of the first content within the second time period based on the output effect evaluation information in the following ways: the cloud-side device determines the pre-push ratio of the first content within the second time period based on the recommendation feature parameters; the cloud-side device determines the reduction ratio of the first content within the second time period based on the output effect evaluation information; and the cloud-side device determines the content push ratio of the first content within the second time period based on the pre-push ratio and the reduction ratio.

[0032] In one possible implementation, the recommendation feature parameters include at least one of the following: user characteristics corresponding to the end device, the popularity of the pre-push content stored in the cloud device, and the value of the pre-push content.

[0033] Secondly, embodiments of this application provide an information recommendation method for a system including a terminal device and a cloud device. The method includes: the terminal device acquiring output effect evaluation information of a first content among the content pushed within a first time period; the terminal device sending the output effect evaluation information to the cloud device; the cloud device determining a content push ratio corresponding to the first content within a second time period based on the output effect evaluation information, the content push ratio being used to characterize the proportion of the first content in the pushed content, and determining the content pushed to the terminal device based on the content push ratio.

[0034] Thirdly, embodiments of this application provide an information recommendation system, including: a terminal device and a cloud device. The information recommendation system is used to execute the information recommendation method mentioned in embodiments of this application, which is performed in cooperation between the terminal device and the cloud device.

[0035] Fourthly, embodiments of this application provide an end-side device, including: one or more processors; one or more memories; the one or more memories storing one or more programs, which, when executed by one or more processors, cause the end-side device to execute the information recommendation method executed by the end-side device in the embodiments of this application.

[0036] Fifthly, embodiments of this application provide a readable storage medium storing instructions, which, when executed on an end-side device, cause the end-side device to perform the information recommendation method mentioned in embodiments of this application. Attached Figure Description

[0037] Figure 1 illustrates some embodiments of this application, showing a scenario in which an information recommendation method is applied;

[0038] Figure 2 shows a comparative schematic diagram of the recommendation page according to some embodiments of this application;

[0039] Figure 3 illustrates an interactive flowchart of an information recommendation method according to some embodiments of this application;

[0040] Figure 4 illustrates an interactive flowchart of an information recommendation method according to some embodiments of this application;

[0041] Figure 5 illustrates an interactive flowchart of an information recommendation method according to some embodiments of this application;

[0042] Figure 6 shows a schematic diagram of an information recommendation system according to some embodiments of this application;

[0043] Figure 7 shows a schematic diagram of the hardware structure of an end-side device according to some embodiments of this application;

[0044] Figure 8 shows a schematic diagram of the hardware structure of a cloud-side device according to some embodiments of this application. Detailed Implementation

[0045] This application provides an information recommendation method, system, end-side device, and storage medium.

[0046] The end-side devices mentioned in the embodiments of this application include, but are not limited to, mobile stations (MS) and mobile terminals (MT). For example, end-side devices can be mobile phones, smart TVs, wearable devices, tablets, desktop computers, laptops, virtual reality (VR) devices, augmented reality (AR) devices, terminals in industrial control, terminals in self-driving vehicles, terminals in remote medical surgery, terminals in smart grids, terminals in transportation safety, terminals in smart cities, terminals in smart homes, and so on. The embodiments of this application do not limit the specific form of the end-side devices.

[0047] The cloud-side devices provided in this application include, but are not limited to, servers, server clusters, and infrastructure and platforms supporting cloud computing services. For example, cloud-side devices can be high-performance physical servers within large data centers, virtual servers created through virtualization technology, and high-availability clusters and distributed storage clusters composed of multiple servers used to provide various cloud service modes such as Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). As long as it can serve as a core component in a cloud computing architecture and support functions such as data processing, storage, distribution, and management, it can be considered a cloud-side device covered by this application. This application does not strictly limit the specific form of the cloud-side device.

[0048] The following describes the application scenario of the information recommendation method mentioned in the embodiments of this application, taking the terminal device as mobile phone 100 and the cloud device as server 200 as an example.

[0049] Figure 1 illustrates a scenario where mobile phone 100 and server 200 perform information recommendation. As shown in Figure 1, when mobile phone 100 detects that a user triggers the recommendation control 1011 in a social app, the social app on mobile phone 100 sends a recommendation request to the recommendation system 201 of the server 200 corresponding to the social app. The recommendation system 201 can filter content from the content information database 202 based on user characteristics, the popularity of the content, or the value of the content, and send the corresponding content to the social app on mobile phone 100. After receiving the content, the social app on mobile phone 100 renders the content and displays it on the recommendation page 101. It is understood that the above recommendation scenario is only an illustrative example, and the information recommendation method mentioned in the embodiments of this application can be used in any scenario that requires information recommendation.

[0050] The content can take various forms (such as images, videos, live streams, music, text posts, etc.). This content can be displayed to users through multiple sub-pages on the recommendation page 101. For example, as shown in Figure 1, recommendation page 101 includes a first sub-page 101-1, a second sub-page 101-2, and a third sub-page 101-3. The first sub-page 101-1 displays articles, the second sub-page 101-2 displays videos, and the third sub-page 101-3 displays live product broadcasts.

[0051] It is understandable that when a device displays or plays content, issues may arise due to factors such as the device's hardware and software performance (e.g., display performance, processor resources) or network conditions. For example, with video content, insufficient processor resources, low screen refresh rates, or poor network connectivity may cause stuttering or playback failures, thus impacting the user experience.

[0052] Therefore, to improve the display or playback effect on the device side and enhance the user experience, this application provides an information recommendation method. The device side can obtain output effect evaluation information (such as display effect evaluation information, playback effect evaluation information, etc.) obtained by evaluating the output effect of preset type content on the device side, and send the output effect evaluation information to the cloud-side device. The cloud-side device can adjust the push ratio of preset type content to the device side based on the output effect evaluation information. For example, if the playback effect evaluation information of video content sent by a device side to the cloud-side device over a period of time indicates that the video content plays poorly on the device side, the cloud-side device can reduce the push ratio of video content when sending content to that device side later. In this way, it is possible to dynamically adjust content recommendations based on different device side devices, adapt to the device side devices, improve the output effect of the device side devices, and enhance the user experience.

[0053] In some embodiments, the output effect evaluation information corresponding to the preset type of content may include an output effect score for the preset type of content within a preset time period. For example, a higher output effect score indicates a better output effect, and a lower output effect score indicates a worse output effect. Alternatively, it can be set so that a lower output effect score indicates a better output effect, and a higher output effect score indicates a worse output effect.

[0054] For example, for video content, a lower output quality score indicates that the video content has a better output quality, meaning that the video may play smoothly without stuttering, being unable to play, or experiencing few stuttering issues. Conversely, a higher output quality score indicates that the video content has a worse output quality, such as experiencing more stuttering or being unable to play.

[0055] In some embodiments, the cloud-side device can determine the reduction ratio of a preset type of content based on the output effect score of the preset type of content within a preset time period. For example, if the total score of video content is 10 points, and the output effect score is between 3 and 5 points, the reduction ratio can be 30%, that is, 30% of the video content is reduced from the original recommendation volume. If the evaluation score is between 8 and 10 points, the reduction ratio can be 100%, that is, video content will no longer be recommended in subsequent recommendations.

[0056] In some embodiments, the cloud-side device can determine a first push ratio (i.e., the push ratio corresponding to the original recommendation volume) for a given type of content based on recommendation feature parameters such as user characteristics (e.g., age, viewing preferences), the popularity of the content to be pushed, or the value of the content. Then, it determines the final push ratio for the corresponding type of content based on a reduction ratio. For example, the cloud-side device determines that the push ratio for video content is 50% based on user characteristics, the popularity of the content, and the value of the content. If the output effect rating of the video content is between 3 and 5 points, the reduction ratio for the video content can be determined to be 30%. In this case, the final push ratio for the video content can be determined to be 20%.

[0057] In some embodiments, the output performance score of the preset content type can be determined based on the scores of each feature parameter of the preset content type and the weights corresponding to each feature parameter. For example, for the preset content type being video content, the feature parameters of the video content type include at least one of the following: first frame playback status, playability status, frame rate status, stuttering status, central processing unit (CPU) usage status, and memory usage status.

[0058] In some embodiments, the weight of each feature parameter can be determined based on the degree of influence of each feature parameter on the output effect of each type of content (i.e., the degree of influence on the user experience). The greater the degree of influence, the greater the weight. For example, for the preset type of content, which is video content, the first frame delay is too large or the video content is unplayable, which has a significant impact on the user experience. That is, the first frame playback status and the playability status have a significant impact on the output effect of the video content, so a larger weight can be set for the first frame playback status and the playability status.

[0059] In some embodiments, the preset content type may include one or more types of content, such as video content, music content, e-book content, image content, and any other type of content. In some embodiments, the content types may be further subdivided; for example, video content may include high frame rate video content, low frame rate video content, etc., but this application embodiment does not impose such limitations.

[0060] In some embodiments, the preset type of content can be content that has high or special requirements for the software, hardware, or network resources of electronic devices. For example, video content generally requires high CPU and memory resources; music content has specific requirements for the audio chip or audio decoding capabilities of electronic devices; and e-book content may have specific requirements for the text formats supported by the player.

[0061] It is understood that the hardware and software performance of electronic devices (such as processor resources, device display performance, etc.) and network resources will affect the rendering effect of the terminal device on the aforementioned preset content types. Therefore, by evaluating the rendering effect of the content in this embodiment, it is possible to recommend content adapted to the electronic device based on its hardware and software performance (processor resources, device display performance, etc.) or network resources. This can improve the output effect of the electronic device and enhance the user experience.

[0062] The following section uses preset content types, including video content, as an example to compare the output effects of different end devices on the content.

[0063] For example, if the display performance of the first device and the second device differs, and a user logs into the same account on both devices, the content recommended on the recommendation interfaces of the first and second devices will differ. For instance, if the display of the first device does not support high frame rate video, while the display of the second device does, then based on the information recommendation method of this application embodiment, the recommendation interface of the first device may not contain high frame rate video content, while the recommendation interface of the second device may contain high frame rate video content.

[0064] For example, if the processor resources of the first and second edge devices differ, and a user logs into the same account on both devices, the content recommended on the recommendation pages of the first and second edge devices will differ. For instance, if the processor performance of the first edge device is lower, video playback will generally experience more stuttering, while the processor performance of the second edge device is higher, video playback will generally not experience stuttering. Therefore, based on the information recommendation method of this application embodiment, the proportion of video type content displayed on the recommendation page of the first edge device will be lower than the proportion displayed on the recommendation page of the second edge device.

[0065] For example, if the first end-side device and the second end-side device are connected to different network devices, and the user logs in to the same account on both devices, the recommended interfaces displayed on the two end-side devices may differ. For instance, if the first end-side device is connected to a first network device with a lower network speed, resulting in more frequent buffering when playing videos; and the second end-side device is connected to a second network device with a higher network speed, resulting in no buffering when playing videos, then based on the information recommendation method of this application embodiment, the proportion of video type content displayed on the recommended page of the first end-side device is lower than the proportion of video type content displayed on the recommended page of the second end-side device.

[0066] Figure 2 illustrates a comparison of the recommendation page when a user logs into the same social APP account on device 300 (with poor network speed) and device 400 (with good network speed).

[0067] As shown in Figure 2, the recommendation page 301 of electronic device 300 includes a first subpage 301-1, a second subpage 301-2, a third subpage 301-3, and a fourth subpage 301-4. The first subpage 301-1 displays post content, the second subpage 301-2 displays image content, the third subpage 301-3 displays image content, and the fourth subpage 301-4 displays article content. Similarly, the recommendation page 401 of electronic device 400 includes a first subpage 401-1, a second subpage 401-2, a third subpage 401-3, and a fourth subpage 401-4. The first subpage 401-1 displays live stream content, the second subpage 401-2 displays video content, the third subpage 401-3 displays video content, and the fourth subpage 401-4 displays article content.

[0068] As can be seen from Figure 2, the proportion of video content displayed on the page of the terminal device 300 with poor network speed is significantly lower than the proportion of video content displayed on the page of the terminal device 400 with good network speed. It is understood that the content of the above-described recommendation page is merely illustrative, and this embodiment does not specifically limit the actual content of the recommendation page.

[0069] Thus, the information recommendation method provided in this application embodiment can dynamically adjust content recommendations based on different terminal devices, adapt to terminal devices, improve the output effect of terminal devices, and enhance the user browsing experience.

[0070] The information recommendation method mentioned in the embodiments of this application will be described in detail below.

[0071] Figure 3 illustrates the interaction flowchart of an information recommendation method. As shown in Figure 3, the information recommendation method includes:

[0072] 101: The terminal device obtains the output effect evaluation information of the preset type of content in the content pushed within the first time period.

[0073] In some embodiments, the duration of the first time period can be determined according to actual needs. For example, the first time period can be 10 minutes, meaning the edge device can evaluate the output effect every 10 minutes. For instance, the first time period can be the duration corresponding to each rendering of the recommendation page, meaning that an output effect evaluation will be performed every time the recommendation page is rendered.

[0074] In some embodiments, the preset type content may refer to the first content mentioned in the embodiments of this application. The preset type content may include one or more types of content, such as video content, music content, e-book content, image content, and any other type of content. In some embodiments, the content types may be further divided. For example, video content may include high frame rate video content, low frame rate video content, etc., but this embodiment of the application does not impose such limitations.

[0075] In some embodiments, when the preset type content includes one type of content, the output effect evaluation information of the preset type content may include the output effect score of that type of content. When the preset type content includes multiple types of content, the output effect evaluation information of the preset type content may include the output effect scores corresponding to each type of content.

[0076] In some embodiments, the output effect score of the preset type of content can be determined by the scores of each feature parameter of the preset type of content within a first time period and the weights corresponding to each feature parameter.

[0077] For example, when the preset content type is video content, the characteristic parameters of video content include at least one of the following: first frame playback status, playability (whether it can be played), frame rate, stuttering, CPU usage, and memory usage; when the preset content type is audio content, the characteristic parameters of audio content include at least one of the following: first frame playback status, playability, stuttering, CPU usage, and memory usage; when the preset content type is e-book content, the characteristic parameters of e-book content include at least one of the following: homepage display status, open status (whether it can be opened), stuttering, CPU usage, and memory usage; when the preset content type is image content, the characteristic parameters of image content include at least one of the following: image loading time, open status, CPU usage, and memory usage.

[0078] In some embodiments, the weight of each feature parameter can be determined based on the degree of influence of each feature parameter on the output effect of each type of content (i.e., the degree of influence on the user experience). The greater the degree of influence, the greater the weight. For example, for the preset type of content, which is video content, the first frame delay is too large or the video content is unplayable, which has a significant impact on the user experience. That is, the first frame playback status and the playability status have a significant impact on the output effect of the video content. Therefore, a larger weight can be set for the first frame playback status and the playability status.

[0079] The scoring of each feature parameter for different types of content and the calculation method of the output effect score corresponding to different types of content will be detailed later, and will not be repeated here.

[0080] In some embodiments, when the edge device detects that an information recommendation application (e.g., a first application) has been launched, it can send a first request to the cloud-side device corresponding to the information recommendation application. The first request is used to obtain evaluation strategy information. After receiving the evaluation strategy information sent by the cloud-side device, the edge device obtains the output effect evaluation information of preset type content within a first time period based on the evaluation strategy information. In this way, the edge device does not need to have the evaluation strategy information built in, saving memory of the edge electronic device.

[0081] In some embodiments, the evaluation strategy information may include the feature parameters of preset content types and the weights corresponding to each feature parameter, the evaluation interval duration (e.g., the duration corresponding to the first time period), the evaluation effective time, and the specific calculation method for the output effect evaluation information. The information recommendation application can refer to any application with information recommendation functionality. The cloud-side device may include the server corresponding to the information recommendation application.

[0082] In some embodiments, the evaluation strategy information can also be pre-stored in the edge device, eliminating the need for the edge device to retrieve it from the cloud device and saving communication resources. When the information recommendation application on the edge device starts, the edge device retrieves the output effect evaluation information of preset types of content from the pushed content within a first time period based on the evaluation strategy information stored in the edge device.

[0083] 102: The end-side device sends output performance evaluation information to the cloud-side device.

[0084] In some embodiments, when the end-side device obtains the output effect evaluation information of the preset type of content within the first time period, it can send the output effect evaluation information of the preset type of content within the first time period to the cloud-side device.

[0085] In some embodiments, when the duration of the first time period is the duration corresponding to each rendering of the recommendation page, the terminal device will perform an output effect information calculation once per rendering, and can send a content recommendation request carrying output effect evaluation information to the cloud device each time it detects that the next page of recommended content has been retrieved. When the duration of the first time period is a preset duration, such as 10 minutes, the terminal device can perform an output effect information calculation every 10 minutes and send output effect evaluation information to the cloud device every 10 minutes.

[0086] 103: The cloud-side device determines the content push ratio of the preset type of content corresponding to the second time period based on the output effect evaluation information, and determines the content to be pushed to the end-side device based on the content push ratio of the preset type of content within the second time period.

[0087] In some embodiments, the second time period can be a subsequent time period of the first time period, and the duration of the second time period can be determined according to actual needs. In some embodiments, the second time period can be the same as the duration of the first time period. For example, the first and second time periods can be 10 minutes, meaning the edge device can evaluate the output effect every 10 minutes, and the application period for each output effect evaluation is also 10 minutes. For example, if the first time period is 9:00-9:10, then the second time period is 9:10-9:20. As another example, both the first and second time periods can be the duration corresponding to each rendering of the recommendation page, meaning that each rendering of the recommendation page will perform an output effect evaluation, and the recommendation ratio corresponding to the output effect evaluation will be applied to the next rendering of the recommendation page. In some embodiments, the second time period can be different from the duration of the first time period; this application embodiment does not specifically limit this.

[0088] In some embodiments, the cloud-side device can determine the pre-push ratio of preset content types within a second time period based on recommended feature parameters, determine the reduction ratio of preset content types within a second time period based on output effect evaluation information, and determine the content push ratio of preset content types within a second time period based on the pre-push ratio and the reduction ratio. It can be understood that the push ratio of preset content types is used to characterize the proportion of preset content types in the pushed content. For example, preset content types include video content, and the push ratio of preset content types is used to characterize the proportion of video content types in the pushed content.

[0089] In some embodiments, the recommended feature parameters include at least one of the following: user characteristics corresponding to the endpoint device, the popularity of the pre-push content stored in the cloud device, and the value of the pre-push content. The user characteristics corresponding to the endpoint device may include parameters reflecting the characteristics of the logged-in account, such as the user's historical viewing habits and the age entered by the user. The popularity of the pre-push content may refer to the predicted click-through rate (CTR) of the content to be pushed stored in the cloud device; a higher predicted CTR indicates higher popularity. The value of the pre-push content may include advertising value, commercial value, etc.

[0090] In some embodiments, determining the reduction ratio of preset type content within a second time period based on the output effect evaluation information of preset type content may include determining the reduction ratio based on the score range of the output effect rating of the preset type content. For example, if the total score of video type content is 10 points, and the output effect rating is between 3 and 5 points, the reduction ratio can be 30%, meaning that 30% of the video type content will be reduced from the original recommendation volume. If the evaluation score is between 8 and 10 points, the reduction ratio is 100%, meaning that video type content will no longer be recommended in the recommendations during the second time period.

[0091] For example, for video content, cloud-based devices can determine that the push ratio of video content is 50% based on user characteristics, the popularity of the content, and the value of the content. If the reduction ratio of video content is determined to be 30% based on the output effect score of video content, then the final push ratio of video content in the second time period can be determined to be 20%.

[0092] In some embodiments, the output effect evaluation information corresponding to the preset type of content sent by the end device within a first time period may include the scores of each feature parameter corresponding to the preset type of content within the first time period. In this embodiment, the cloud device may determine the output effect score corresponding to the preset type of content based on the scores of each feature parameter corresponding to the preset type of content and the weights of each feature parameter corresponding to the preset type of content, and determine the content push ratio within a second time period based on the output effect score corresponding to the preset type of content.

[0093] In some embodiments, the output effect evaluation information corresponding to the preset type of content within a first time period sent by the end device may include the sampled values ​​and preset standard values ​​of each feature parameter corresponding to the preset type of content within the first time period. In this embodiment, the cloud device may calculate the score of each feature parameter of the preset type of content based on the sampled values ​​and preset standard values ​​of each feature parameter corresponding to the preset type of content, and determine the output effect score corresponding to the preset type of content based on the score of each feature parameter of the preset type of content and the weight of each feature parameter corresponding to the preset type of content, and determine the content push ratio of the preset type of content within a second time period based on the output effect score corresponding to the preset type of content.

[0094] In some embodiments, after determining the content push ratio of a preset content type within a second time period, the cloud-side device can, each time it receives a content recommendation request from the edge device, determine the content to be pushed to the edge device based on the preset content push ratio and recommendation feature parameters. For example, if it is determined that the final push ratio of video content in the second time period is 20%, then the cloud-side device can determine the content to be pushed, which contains 20% video content, based on the recommendation feature parameters within the second time period.

[0095] It is understandable that when the device on the client side detects that the user is pulling the next page of recommendations, it can send a content recommendation request to the cloud-side device.

[0096] 104: Cloud-side devices push corresponding content to end-side devices.

[0097] In some embodiments, each time the cloud-side device receives a recommendation request from the end-side device, it can push the corresponding content to the end-side device if it determines the content to be pushed to the end-side device, so that the end-side device can display the content pushed by the cloud-side device.

[0098] Thus, based on the information recommendation method provided in this application embodiment, the playback or display of preset type content on the terminal device can be accurately evaluated based on the output effect information of preset type content, thereby adjusting the push ratio of preset type content. In other words, it can realize dynamic adjustment of content recommendation based on different terminal devices, adapt to terminal devices, improve the output effect of terminal devices, and enhance the user's experience of browsing recommendation pages.

[0099] The following example, using video content as the preset content type, details the implementation method for obtaining the output effect score of the preset content type within the first time period mentioned in step 101.

[0100] In some embodiments, the video content within the first time period may include one or more video contents.

[0101] If the video content in the first time period includes a video, then the scores of each feature parameter of that video can be used as the scores of each feature parameter of the video content in the first time period.

[0102] When the video content within the first time period includes multiple video contents, the score of each feature parameter of the video content within the first time period can be determined based on the scores of each feature parameter of the multiple video contents. For example, the average score of each feature parameter of the multiple video contents can be used as the score of each feature parameter of the video content within the first time period, or the highest and lowest scores among the scores corresponding to the multiple video contents can be removed, and the average score of each feature parameter of the remaining video contents can be used as the score of each feature parameter of the video content within the first time period, etc. The embodiments of this application do not limit this.

[0103] For example, for any feature parameter of video content, the score of each video content for that feature parameter can be determined based on the ratio of the difference between the sampled value of that feature parameter of each video content and the preset standard value within the first time period to the preset standard value, and the average score of each video content can be used as the score of that feature parameter of the video content within the first time period.

[0104] For example, for any feature parameter of video content, the score of that feature parameter for the video content in the first time period can be determined based on the ratio of the difference between the average of the sampled values ​​of that feature parameter for each video content in the first time period and the preset standard value to the preset standard value. For example, the method for determining the score of each feature parameter can be as shown in Formula 1 below:

[0105] in, This represents the average value of the sampled values ​​corresponding to the feature parameters, i.e. m represents the number of samples, such as the number of video contents in the first time period, S(i) is the preset standard value of the feature parameter, and Score(i) is the score corresponding to the feature parameter.

[0106] In some embodiments, when a user browses a recommendation page (e.g., when browsing the content of the recommendation page), the edge device can collect the sampled values ​​corresponding to each feature parameter of each content in the preset type of content in real time.

[0107] The following section uses 10 video content items within the first time period as an example to introduce the scoring method for each feature parameter of video content.

[0108] (1) The scoring method for obtaining the first frame playback status of video content within the first time period may include:

[0109] Based on the sampling difference between the first frame display time of each of the 10 video contents within the first time period and the time when the terminal device detects the first operation that triggers the video content playback (e.g., the time when the user clicks the play button to trigger video content playback, or the time when the user swipes the recommendation page to automatically trigger video content playback, etc.), a score for the first frame playback start status of the corresponding video content type is determined. Specifically, the method for determining the score for the first frame playback start status of the corresponding video content type can be as shown in Formula 2.

[0110] Wherein, Score(1) is the score for the first frame playback of video content. S(1) is the average of the sampling differences between the first frame display time and the time when the terminal device detects the first operation that triggers the playback of the video content for each of the 10 video contents. S(1) is the preset standard deviation between the first frame display time and the time when the terminal device detects the first operation that triggers the playback of the video content.

[0111] For example, when If the time interval is 5ms and the time interval for S(1) is 2ms, then the score for Score(1) is 1.5 points.

[0112] In some embodiments, the method for calculating the sampling difference between the first frame display time and the time when the terminal device detects the corresponding first operation triggering video content playback can be as follows: the time when the user clicks the play button or the play button is automatically triggered is recorded by marking points, and the time when the user clicks the play button or the play button is automatically triggered is taken as the time when the first operation triggers video content playback. In addition, the start time of the video player can be monitored and taken as the first frame display time. The sampling difference between the first frame display time and the time when the terminal device detects the corresponding first operation triggering video content playback is calculated in the event callback.

[0113] (2) Scoring methods for obtaining the frame rate of video content within the first time period may include:

[0114] Based on the sampling difference between the frame rate of 10 video content items within the first time period and the screen refresh rate of the terminal device, a score is determined for the frame rate of the video content type. Specifically, the method for determining the frame rate score for the video content type is as shown in Formula 3.

[0115] Wherein, Score(2) is the score for the frame rate of the video content. S(2) is the average of the sampling differences between the frame rate corresponding to 10 video contents and the screen refresh rate of the terminal device, and S(2) is the preset standard deviation between the frame rate corresponding to the video contents and the screen refresh rate of the terminal device.

[0116] For example, If the frame rate is 120fps and the frame rate of S(2) is 60fps, then the score of Score(2) is 1 point.

[0117] In some embodiments, the frame rate of a video can be obtained through an application programming interface (API) such as the MediaMetadataRetriever function or the AVMetadataExtractor function, which retrieves metadata from the video. The screen refresh rate of the device can then be obtained through the read-only memory (ROM) of the device.

[0118] (3) The scoring methods for obtaining the CPU usage of video content within the first time period may include:

[0119] Based on the ratio of CPU resources used by the endpoint device during the playback of 10 corresponding video contents to the total CPU resources of the endpoint device within the first time period, a score is determined for the CPU usage of each video content type. Specifically, the method for determining the ratio of CPU resources used by each video content during playback to the total CPU resources of the endpoint device can be as follows: during video playback, the CPU ratio used by the video playback process is sampled periodically, and the highest CPU ratio among the video playback processes is taken as the ratio of CPU resources used by each video content during playback to the total CPU resources of the endpoint device.

[0120] The scoring method for determining the CPU usage corresponding to different video content types can be shown in Formula 4.

[0121] Wherein, Score(3) is a score for the CPU usage of video content. S(3) is the average ratio of CPU resources occupied by 10 video content playback to the total CPU resources of the terminal device, and S(3) is the preset standard value of the ratio of CPU resources occupied by video content playback to the total CPU resources of the terminal device.

[0122] For example, S is 30%, and S(3) is 20%. Therefore, Score(3) is 0.5 points.

[0123] (4) The scoring methods for obtaining the memory usage of video content within the first time period may include:

[0124] Based on the ratio of memory resources occupied by the 10 video contents detected by the edge device during the first time period to the total memory resources of the edge device, a score is determined for the memory usage of each video content. The ratio of memory resources occupied by each video content to the total memory resources of the edge device can be determined in two ways: First, the memory ratio occupied by the video playback process is sampled periodically during playback, and the highest memory ratio among the CPU ratios of the video playback process is used as the ratio of memory resources occupied by each video content to the total memory resources of the edge device. Second, the average memory ratio occupied during playback is also sampled periodically and used as the ratio of memory resources occupied by each video content to the total memory resources of the edge device.

[0125] The scoring method for determining the memory usage corresponding to different video content types can be shown in Formula 5:

[0126] Where Score(4) is the score for the memory usage of video content. S(4) is the average ratio of the memory resources occupied by 10 video contents during playback to the total memory resources of the terminal device, and S(4) is the preset standard value of the ratio of the memory resources occupied by video contents during playback to the total memory resources of the terminal device.

[0127] For example, The score is 30%, and S(4) is 10%. Therefore, Score(4) is 2 points.

[0128] (5) The scoring methods for obtaining the stuttering status of video content within the first time period may include:

[0129] Based on the number of stutters detected by the edge device during the playback of 10 corresponding video content items within the first time period, a score for the stuttering level of each video content type is determined. Specifically, the method for determining the stuttering level score for each video content type is shown in Formula 6.

[0130] Score(5) is a rating of the stuttering or buffering issues in video content. S(5) is the average number of times the playback state changes to the stopped state during the playback of 10 video contents. S(5) is the preset standard value of the number of times the playback state changes to the stopped state during the playback of video contents.

[0131] For example, If the value is 5, then S(5) is 2. Therefore, Score(5) is 1.5 points.

[0132] In some embodiments, the system can listen for video player state change notification events and record a change from a playing state to a stopped state as a stutter. In some embodiments, when the player has two states—playback complete and stopped—a change from a playing state to a stopped state can be recorded as a stutter. In some embodiments, when the player has both a playback complete and a stopped state, such as both being stopped, the actual state can be distinguished by the video playback progress. For example, if the video playback progress is 100%, the current state is determined to be a playback complete state. If the video playback progress is less than 100%, the current state is determined to be a stopped state.

[0133] (6) The scoring methods for obtaining the playability of video content within the first time period may include:

[0134] Based on the playback status of 10 video content items detected by the edge device within the first time period, a score is determined for the playability of each video content type. Specifically, a sample value of 1 corresponds to a playable video content, and a sample value of 2 corresponds to an unplayable video content. The method for determining the playability score for each video content type is shown in Formula 7.

[0135] Wherein, Score(6) is a score for the playability of video content. S(6) is the average of the sampled values ​​corresponding to the playback status of 10 video contents, and S(6) is the preset standard value corresponding to the playback status of the video contents.

[0136] For example, Score(5) is 2.5, and S(5) is 1. Therefore, Score(5) is 1.5 points.

[0137] In some embodiments, key events such as error (onError) events can be listened to when the video player is playing a video. When an onError event is detected, it can be determined that the current video is not playable on the current device. For example, the onError event may be triggered by situations such as the current device not supporting the video format. When no onError event is detected, it can be determined that the current video is playable on the current device.

[0138] In some embodiments, when the content of the video type within the first time period includes a video content, for example, including a first video content, the scores of each feature parameter of the video type content are obtained, including at least one of the following:

[0139] Based on the difference between the display time of the first frame corresponding to the first video content within the first time period and the time when the terminal device detects the first operation that triggers the playback of the first video content, a score for the start-up status of the first frame corresponding to the first video content is determined; the score for the start-up status of the first frame corresponding to the first video content is used as the score for the start-up status of the first frame of the video type content within the first time period.

[0140] Based on whether the first video content was successfully played within the first time period, a score is obtained for the playability of the first video content; the score for the playability of the first video content is used as the score for the playability of video type content within the first time period.

[0141] The frame rate score of the first video content is determined based on the difference between the frame rate of the first video content and the refresh rate of the display screen of the terminal device within the first time period; the frame rate score of the first video content is used as the frame rate score of the video type content.

[0142] Based on the number of times the first video content stutters within the first time period, a score for the stuttering situation of the first video content is obtained; the score for the stuttering situation of the first video content is used as the score for the stuttering situation of video content of the same type within the first time period.

[0143] The CPU usage score of the first video content is determined based on the ratio of CPU resources used during the playback of the first video content to the total CPU resources of the terminal device within the first time period; the CPU usage score corresponding to the first video content is used as the CPU usage score of the video type content within the first time period.

[0144] The memory usage score of the first video content is determined based on the ratio of memory resources used during playback of the first video content to the total memory resources of the device within the first time period. This memory usage score is then used as the overall memory usage score for video content within the first time period.

[0145] After obtaining the scores of each feature parameter of the video content within the first time period, the output effect score of the video content within the first time period can be determined based on the scores and weights of each feature parameter. Specifically, the output effect score of the video content within the first time period can be determined using the following formula 8:

[0146] Where Score(i) is the score of each of the aforementioned feature parameters, W(i) represents the weight of the feature parameter, n represents the number of content feature parameters, and RAT is the output effect score of video type content in the first time period.

[0147] For example, if the scores for the video type's feature parameters—first frame playback status, frame rate, CPU usage, memory usage, stuttering, and playability—are 1.5, 1, 0.5, 2, 1.5, and 1.5 respectively, and their weights are 20%, 10%, 5%, 5%, 10%, and 50% respectively, then the final calculated display score for the video type content is 1.425.

[0148] It is understood that in some embodiments, the cloud-side device may include an operation management system, a recommendation system, and an information flow server. The operation management system and the recommendation system may refer to different servers, or they may be different functional units within different servers, or different functional units within the same server, etc., and this application does not limit the scope of these embodiments. In some embodiments, the operation management system may be used to configure evaluation strategy information, push ratio strategy parameters, etc., and may be used to send evaluation strategy information to the information flow server and send push ratio strategy parameters to the recommendation system; the recommendation system may be used to determine the push content for a second time period based on the push ratio strategy parameters, and push the corresponding content to the information flow server within the second time period; the information flow server may be used to send evaluation strategy information to the edge device, and send the output effect evaluation information received from the edge device to the recommendation system, and may also be used to push the corresponding content to the edge device.

[0149] Figure 4 illustrates a flowchart of an information recommendation method, using cloud-based devices including an operation management system, a recommendation system, and an information flow server (where the operation management system and recommendation system refer to different servers). As shown in Figure 4, the method includes:

[0150] 201: The operations management system sends evaluation strategy information to the information flow server and push ratio strategy parameters to the recommendation system.

[0151] In some embodiments, operators can pre-configure evaluation strategy information and push ratio strategy parameters in the operation management system. The operation management system can send the evaluation strategy information to the information flow server and the push ratio strategy parameters to the recommendation system.

[0152] In some embodiments, the evaluation strategy information may include various feature parameters of preset content types, the weights corresponding to each feature parameter, the evaluation interval duration (e.g., the duration corresponding to the first time period), the evaluation effective time (e.g., the second time period), and the specific calculation method for the output effect evaluation information. For example, it may include various feature parameters of video content types, the weights corresponding to each feature parameter, the evaluation interval duration (e.g., 10 minutes), the evaluation effective time (e.g., 10 minutes), and the specific calculation method for the output effect evaluation information.

[0153] The push ratio strategy parameters can include: the push reduction ratio corresponding to different score ranges for preset content types. For example, for video content, the reduction ratio can be 30% when the output effect score is 3-5 points, and 100% when the evaluation score is 8-10 points.

[0154] The push ratio strategy parameters may also include recommendation feature parameters for determining the pre-push ratio of preset content types, as well as specific methods for determining the pre-push ratio based on the recommendation feature parameters, and methods for determining the content to be pushed in the second time period based on the recommendation feature parameters and the recommendation ratio of preset content types.

[0155] It is understandable that the parameters in the evaluation strategy information and push ratio strategy parameters can change dynamically. For example, operations managers can adjust the parameters in the evaluation strategy information and push ratio strategy parameters in real time according to actual needs.

[0156] 202: The terminal device detected information recommending the launch of the APP.

[0157] In some embodiments, the information recommendation app can refer to any application with information recommendation functionality. The device can confirm the launch of the information recommendation app when it detects the startup of the thread corresponding to the information recommendation app. The device can detect the launch of the information recommendation app in any feasible manner. This application does not limit this approach.

[0158] 203: The end device sends a first request to the information flow server. The first request is used to obtain evaluation strategy information.

[0159] In some embodiments, when the edge device detects that the information recommendation app has been launched, it may send a first request to the server corresponding to the information recommendation app, i.e., the information flow server. The first request is used to obtain evaluation strategy information.

[0160] 204: The information flow server sends evaluation strategy information to the end-side device.

[0161] In some embodiments, upon receiving a first request, the information flow server may send evaluation policy information to the end-side device.

[0162] 205: The end-side device evaluates the preset type of content in the first time period based on the evaluation strategy information and obtains the output effect evaluation information.

[0163] In some embodiments, the edge device evaluates the preset type of content in the first time period based on the evaluation strategy information, and the method for obtaining the output effect score can be as described in 101 above, which will not be repeated here.

[0164] 206: The end device sends output effect evaluation information to the information flow server.

[0165] In some embodiments, when the duration of the first time period is the duration corresponding to each rendering of the recommendation page, the terminal device will perform an output effect information calculation once per rendering, and can send a content recommendation request carrying output effect evaluation information to the cloud device each time it detects that the next page of recommended content has been retrieved. When the duration of the first time period is a preset duration, such as 10 minutes, the terminal device can perform an output effect information calculation every 10 minutes and send output effect evaluation information to the cloud device every 10 minutes.

[0166] 207: The information flow server sends output effect evaluation information to the recommendation system.

[0167] In some embodiments, after the information flow server obtains the output effect evaluation information sent by the end device, it sends the output effect evaluation information to the recommendation system.

[0168] 208: The recommendation system determines the content to be pushed in the second time period based on the push ratio strategy parameters.

[0169] In some embodiments, the recommendation system can determine the push ratio of preset content types in the second time period based on push ratio strategy parameters and output effect evaluation information of preset content types in the first time period using a preset algorithm (e.g., Content2Display). It can then determine the content to be pushed to the edge device in the second time period based on the push ratio of the preset content types and recommendation feature parameters. For example, if the final push ratio of video content in the second time period is determined to be 20%, then the system can determine the content to be pushed in the second time period based on the recommendation feature parameters, which will contain 20% video content.

[0170] The detailed method for determining the push content for the second time period based on the push ratio strategy parameters can be found in step 103 above, and will not be repeated here.

[0171] 209: The recommendation system pushes the corresponding content to the information flow server.

[0172] In some embodiments, after determining the content to be pushed, the recommendation system can push the corresponding content to the information flow server.

[0173] 210: The information flow server pushes the corresponding content to the end device.

[0174] In some embodiments, when the information flow server receives push content sent by the recommendation system, it can send the corresponding push content to the end device.

[0175] Thus, based on the information recommendation method provided in this application embodiment, the playback or display of preset type content on the terminal device can be accurately evaluated based on the output effect information of preset type content, thereby adjusting the push ratio of preset type content. In other words, it can realize dynamic adjustment of content recommendation based on different terminal devices, adapt to terminal devices, improve the output effect of terminal devices, and enhance the user's experience of browsing recommendation pages.

[0176] In some embodiments, the cloud-side device may include a recommendation system and an information flow server, wherein the recommendation system may refer to the server or a functional unit deployed within the server. In some embodiments, push ratio strategy parameters can be configured directly in the recommendation system, and evaluation strategy information can be configured in the edge device. Thus, the edge device can determine the output effect evaluation information based on the built-in evaluation strategy information, without needing to obtain the evaluation strategy information from the cloud-side device. The recommendation system does not need to retrieve the push ratio strategy parameters from the operation management system. The recommendation system can be used to determine the push content for the second time period based on the built-in push ratio strategy parameters, and can be used to push the corresponding content to the information flow server; the information flow server is used to send the received output effect evaluation information to the recommendation system, and can be used to push the received content to the edge device.

[0177] Figure 5 illustrates a flowchart of an information recommendation method, using cloud-based devices including a recommendation system and an information flow server as an example. As shown in Figure 5, the method includes:

[0178] 301: The recommendation ratio strategy parameter configured in the recommendation system.

[0179] In some embodiments, push ratio strategy parameters can be pre-configured in the recommendation system and stored in the content information database of the recommendation system.

[0180] In some embodiments, the push ratio strategy parameter may include: a preset reduction ratio for pushes corresponding to different score ranges of content types. For example, for video content, the reduction ratio may be 30% when the output effect score is 3-5 points, and 100% when the evaluation score is 8-10 points.

[0181] The push ratio strategy parameters may also include recommendation feature parameters for determining the pre-push ratio of preset content types, as well as specific methods for determining the pre-push ratio based on the recommendation feature parameters, and methods for determining the content to be pushed in the second time period based on the recommendation feature parameters and the recommendation ratio of preset content types.

[0182] 302: The terminal device detected information recommending the launch of the APP.

[0183] In some embodiments, the information recommendation app can refer to any application with information recommendation functionality. The device can confirm the launch of the information recommendation app when it detects the startup of the thread corresponding to the information recommendation app. The device can detect the launch of the information recommendation app in any feasible manner. This application does not limit this approach.

[0184] 303: The end-side device evaluates the preset type of content in the first time period based on the evaluation strategy information and obtains the output effect evaluation information.

[0185] In some embodiments, the edge device can evaluate preset type content in the first time period based on built-in evaluation strategy information to obtain an output effect score. The method for evaluating the preset type content in the first time period based on the built-in evaluation strategy information to obtain the output effect score can be as described in section 101 above, and will not be repeated here.

[0186] 304: The end device sends output performance evaluation information to the information flow server.

[0187] In some embodiments, when the duration of the first time period is the duration corresponding to each rendering of the recommendation page, the terminal device will perform an output effect information calculation once per rendering, and can send a content recommendation request carrying output effect evaluation information to the cloud device each time it detects that the next page of recommended content has been retrieved. When the duration of the first time period is a preset duration, such as 10 minutes, the terminal device can perform an output effect information calculation every 10 minutes and send output effect evaluation information to the cloud device every 10 minutes.

[0188] 305: The information flow server sends output effect evaluation information to the recommendation system.

[0189] In some embodiments, when the end device evaluates the preset type of content in the first time period based on the evaluation strategy information and obtains the output effect evaluation information, it sends the output effect evaluation information to the information flow server.

[0190] 306: The recommendation system determines the content to be pushed in the second time period based on the push ratio strategy parameters.

[0191] In some embodiments, the recommendation system can determine the push ratio of preset content types in the second time period based on built-in push ratio strategy parameters and output effect evaluation information of preset content types in the first time period using a preset algorithm (e.g., the Content2Display algorithm). It can then determine the content to be pushed to the edge device in the second time period based on the push ratio of the preset content types and recommendation feature parameters. For example, if the final push ratio of video content in the second time period is determined to be 20%, then the system can determine the content to be pushed in the second time period based on the recommendation feature parameters, which will contain 20% video content.

[0192] The detailed method for determining the push content for the second time period based on the push ratio strategy parameters can be found in step 103 above, and will not be repeated here.

[0193] 307: The recommendation system pushes the corresponding content to the information flow server.

[0194] In some embodiments, after determining the content to be pushed, the recommendation system can push the corresponding content to the information flow server.

[0195] 308: The information flow server pushes the corresponding content to the end device.

[0196] In some embodiments, when the information flow server receives push content sent by the recommendation system, it can send the corresponding push content to the end device.

[0197] Thus, based on the information recommendation method provided in this application embodiment, the playback or display of preset type content on the terminal device can be accurately evaluated based on the output effect information of preset type content, thereby adjusting the push ratio of preset type content. In other words, it can realize dynamic adjustment of content recommendation based on different terminal devices, adapt to terminal devices, improve the output effect of terminal devices, and enhance the user's experience of browsing recommendation pages.

[0198] The information recommendation system according to an embodiment of this application is described below. Figure 6 is a schematic diagram of an information recommendation system according to an embodiment of this application. As shown in Figure 6, the information recommendation system includes a terminal device 100 and a cloud device 200.

[0199] The end-side device 100 may include an application layer, an application framework layer, system libraries, and kernel libraries.

[0200] The application layer can include a series of application packages. These application packages can include information recommendation apps, as well as various other types of apps. Information recommendation apps can be any app with recommendation functionality, such as social media apps, shopping apps, etc.

[0201] The information recommendation app includes a content display module and a content rendering effect evaluation module.

[0202] The content display module can be used to showcase the recommendation pages of information recommendation apps. For example, it can display content such as videos, live streams, articles, and posts.

[0203] The content rendering effect evaluation module can be used to obtain output effect evaluation information for preset content types within a first time period. For example, the content rendering effect evaluation module can obtain the scores of feature parameters of different content types within the preset time period and the corresponding weights of the feature parameters. Then, based on the scores of the feature parameters of different content types and the corresponding weights of the feature parameters, the display score of the corresponding content type is obtained.

[0204] The content rendering effect evaluation module includes a feature parameter acquisition module, an evaluation strategy execution module, and a rendering effect score reporting module. The evaluation strategy execution module includes a feature parameter scoring module and a feature parameter weight calculation module.

[0205] Feature parameter acquisition module: can be used to obtain sampled values ​​(or sampled differences) of various feature parameters of preset type content.

[0206] Feature parameter scoring module: It can be used to obtain the score of feature parameters of preset type content. For example, it can be used to obtain the score of preset type content within a first time period.

[0207] Feature parameter weight calculation module: It can be used to calculate the output effect score based on the feature parameters and the corresponding weights of the preset content type.

[0208] Content rating and reporting module: This module can be used to send output performance ratings to cloud-based devices.

[0209] The application framework layer provides APIs and a programming framework for applications within the application layer. It includes predefined functions and may include a window manager, content provider, view system, resource manager, notification manager, etc.

[0210] The window manager is used to manage windowed applications. It can retrieve screen size, determine the presence of a status bar, lock the screen, and capture screenshots, among other things.

[0211] Content providers store and retrieve data, making that data accessible to applications. This data can include videos, images, audio, phone calls made and received, browsing history and bookmarks, phone books, and more.

[0212] A view system includes visual controls, such as controls for displaying text and controls for displaying images. View systems can be used to build applications. A display interface can consist of one or more views. For example, a display interface including a text notification icon could include views for displaying text and views for displaying images.

[0213] The file explorer provides applications with various resources, such as localized strings, icons, images, layout files, video files, and more.

[0214] The notification manager allows applications to display notifications in the status bar. These notifications can be used to deliver informational messages and can disappear automatically after a short pause, requiring no user interaction. For example, the notification manager can be used to notify users of completed downloads or message alerts. The notification manager can also display notifications as icons or scrolling text in the top status bar, such as notifications from background applications, or as dialog boxes on the screen. Examples include displaying text messages in the status bar, emitting sounds, vibrating electronic devices, and flashing indicator lights.

[0215] The Android Runtime consists of core libraries and a virtual machine, and is responsible for scheduling and managing the Android system.

[0216] The core library consists of two parts: one part is the functionalities that need to be called by the Java language, and the other part is the Android core library.

[0217] System libraries can include multiple functional modules. For example: surface manager, media libraries, 3D graphics processing libraries (e.g., OpenGL ES), 2D graphics engines, etc.

[0218] The kernel layer is the layer between hardware and software. The kernel layer contains at least display drivers, audio drivers, and sensor drivers.

[0219] The cloud-side device 200 may include an operation management system, a recommendation system, and an information flow server. The operation management system and recommendation system may refer to different servers, or different functional units within different servers, or different functional units within the same server, etc., and this application embodiment does not impose such limitations.

[0220] Operations Management System: It can be used to configure evaluation strategy information, push ratio strategy parameters, etc., and can also be used to send evaluation strategy information to the information flow server and send push ratio strategy parameters to the recommendation system.

[0221] The recommendation system can be used to determine the content to be pushed based on push ratio strategy parameters and push the corresponding content to the information flow server. In some embodiments, when "recommendation system" refers to the server, the recommendation system may include a content information library, which can be used to store and provide content information. In some embodiments, when "recommendation system" refers to a functional unit in the server, the recommendation system and the content information library may reside on the same server.

[0222] The information flow server can be used to send evaluation strategy information to the end-side device, and send the output effect evaluation information received from the end-side device to the recommendation system. It can also be used to push corresponding content to the end-side device.

[0223] The hardware structure of the terminal device is described below, taking mobile phone 100 as an example. As shown in Figure 7, mobile phone 100 may include processor 110, power module 140, memory 180, mobile communication module 130, wireless communication module 120, sensor module 190, audio module 150, camera 170, interface module 160, buttons 103, and display screen 102, etc.

[0224] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on the mobile phone 100. In other embodiments of this application, the mobile phone 100 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0225] Processor 110 may include one or more processing units, such as processing modules or circuits of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Digital Signal Processor (DSP), Micro-programmed Control Unit (MCU), Artificial Intelligence (AI) processor, or Field Programmable Gate Array (FPGA). Different processing units may be independent devices or integrated into one or more processors. Processor 110 may include storage units for storing instructions and data. In some embodiments, the storage unit in processor 110 is a cache memory 180. Processor 100 may be used to execute the methods performed by the end-side device 100 in the embodiments of this application.

[0226] The power module 140 may include a power supply, a power management component, etc. The power supply may be a battery. The power management component manages the charging of the power supply and the power supply to other modules. In some embodiments, the power management component includes a charging management module and a power management module. The charging management module receives charging input from a charger; the power management module connects to the power supply and the processor 110. The power management module receives input from the power supply and / or the charging management module to supply power to the processor 110, the display 102, the camera 170, and the wireless communication module 120, etc.

[0227] The display screen 102 is used to display the recommendation page mentioned in the embodiments of this application.

[0228] The hardware structure of the cloud-side device is described below using server 200 as an example. As shown in Figure 8, server 200 may include one or more processors 1404, system control logic 1408 connected to at least one of the processors 1404, system memory 1412 connected to system control logic 1408, non-volatile memory (NVM) 1416 connected to system control logic 1408, and network interface 1420 connected to system control logic 1408.

[0229] In some embodiments, processor 1404 may include one or more single-core or multi-core processors. In some embodiments, processor 1404 may include any combination of general-purpose processors and special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In embodiments where server 200 employs an Evolved Node B (ENB) or Radio Access Network (RAN) controller 102, processor 1404 may be configured to perform the methods executed by the cloud-side devices in the embodiments of this application.

[0230] In some embodiments, system control logic 1408 may include any suitable interface controller to provide any suitable interface to at least one of the processors 1404 and / or any suitable device or component communicating with system control logic 1408.

[0231] In some embodiments, system control logic 1408 may include one or more memory controllers to provide an interface to system memory 1412. System memory 1412 may be used to load and store data and / or instructions. In some embodiments, the memory 1412 of server 200 may include any suitable volatile memory.

[0232] NVM / memory 1416 may include one or more tangible, non-transitory computer-readable media for storing data and / or instructions. In some embodiments, NVM / memory 1416 may include any suitable non-volatile memory such as flash memory and / or any suitable non-volatile storage device, such as at least one of a hard disk drive (HDD), a compact disc (CD) drive, and a digital versatile disc (DVD) drive.

[0233] NVM / Storage 1416 may include a portion of the storage resources on the device on which Server 200 is installed, or it may be accessible by the device, but is not necessarily part of the device. For example, NVM / Storage 1416 may be accessed over a network via Network Interface 1420.

[0234] Specifically, system memory 1412 and NVM / memory 1416 may respectively include a temporary copy and a permanent copy of instruction 1424. In some embodiments, instruction 1424, hardware, firmware and / or its software components may additionally / alternatively reside in system control logic 1408, network interface 1420 and / or processor 1404.

[0235] Network interface 1420 may include a transceiver for providing a radio interface to server 200, thereby enabling communication with any other suitable device (such as a front-end module, antenna, etc.) over one or more networks. In some embodiments, network interface 1420 may be integrated into other components of server 200. For example, network interface 1420 may be integrated into at least one of processor 1404, system memory 1412, NVM / memory 1416, and firmware device (not shown) with instructions.

[0236] Server 200 may further include: input / output (I / O) device 1432. I / O device 1432 may include a user interface that enables a user to interact with server 200; the design of the peripheral component interface enables peripheral components to also interact with server 200.

[0237] This application provides an information recommendation method for a system including a terminal device and a cloud device. The method includes: the terminal device acquiring output effect evaluation information of a first content among the content pushed within a first time period; the terminal device sending the output effect evaluation information to the cloud device; the cloud device determining the content push ratio corresponding to the first content within a second time period based on the output effect evaluation information, the content push ratio being used to characterize the proportion of the first content in the pushed content, and determining the content pushed to the terminal device based on the content push ratio.

[0238] This application provides an information recommendation system, including: a terminal device and a cloud device. The information recommendation system is used to execute the information recommendation method mentioned in this application, which is performed in cooperation between the terminal device and the cloud device.

[0239] This application provides an end-side device, including: one or more processors; one or more memories; the one or more memories storing one or more programs, which, when executed by one or more processors, cause the electronic device to perform the information recommendation method executed by the end-side device in this application embodiment.

[0240] This application provides a readable storage medium storing instructions. When these instructions are executed on an end-side device, the end-side device performs the information recommendation method mentioned in this application.

[0241] The embodiments disclosed in this application can be implemented in hardware, software, firmware, or a combination of these implementation methods. Embodiments of this application can be implemented as computer programs or program code executable on a programmable system, the programmable system including at least one processor, a storage system (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device.

[0242] Program code can be applied to input instructions to execute the functions described in this application and generate output information. The output information can be applied to one or more output devices in a known manner. For the purposes of this application, the processing system includes any system having a processor such as, for example, a digital signal processor (DSP), a microcontroller, an application-specific integrated circuit (ASIC), or a microprocessor.

[0243] The program code can be implemented using a high-level procedural language or an object-oriented programming language to communicate with the processing system. Assembly language or machine language can also be used when needed. In fact, the mechanisms described in this application are not limited to any particular programming language. In either case, the language can be a compiled language or an interpreted language.

[0244] In some cases, the disclosed embodiments may be implemented in hardware, firmware, software, or any combination thereof. The disclosed embodiments may also be implemented as instructions carried or stored thereon on one or more temporary or non-temporary machine-readable (e.g., computer-readable) storage media, which may be read and executed by one or more processors. For example, the instructions may be distributed via a network or through other computer-readable media. Therefore, machine-readable media may include any mechanism for storing or transmitting information in a machine-readable (e.g., computer-readable) form, including but not limited to floppy disks, optical disks, optical discs, magneto-optical disks, magnetic cards or optical cards, flash memory, or tangible machine-readable storage for transmitting information (e.g., carrier waves, infrared signals, digital signals, etc.) using the Internet in the form of electrical, optical, acoustic, or other propagation signals. Therefore, machine-readable media include any type of machine-readable medium suitable for storing or transmitting electronic instructions or information in a machine-readable (e.g., computer-readable) form.

[0245] In the accompanying drawings, some structural or methodological features may be shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order may not be necessary. Rather, in some embodiments, these features may be arranged in a manner and / or order different from that shown in the illustrative drawings. Furthermore, the inclusion of structural or methodological features in a particular figure does not imply that such features are required in all embodiments, and in some embodiments, these features may be omitted or may be combined with other features.

[0246] It should be noted that all units / modules mentioned in the device embodiments of this application are logical units / modules. Physically, a logical unit / module can be a physical unit / module, a part of a physical unit / module, or a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important factor; the combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed in this application. Furthermore, to highlight the innovative aspects of this application, the above-described device embodiments of this application have not introduced units / modules that are not closely related to solving the technical problems proposed in this application. This does not mean that the above-described device embodiments do not contain other units / modules.

[0247] It should be noted that in the examples and description of this patent, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0248] Although this application has been illustrated and described with reference to certain preferred embodiments thereof, those skilled in the art will understand that various changes in form and detail may be made thereto without departing from the scope of this application.

Claims

1. An information recommendation method, characterized in that, For end-side devices, the method includes: Obtain the output performance evaluation information of the first content among the content pushed within the first time period; Send the output effect evaluation information to the cloud-side device; The cloud-side device receives content pushed during the second time period. The content push ratio of the first content among the content pushed by the cloud-side device during the second time period is determined by the cloud-side device based on the output effect evaluation information. The content push ratio is used to characterize the proportion of the first content in the pushed content.

2. The method according to claim 1, characterized in that, The first content includes multiple types of content, and the output effect evaluation information of the first content includes the output effect scores corresponding to each of the multiple types of content; The method by which the cloud-side device determines the content push ratio of the first content within the second time period based on the output effect evaluation information includes: The cloud-side device determines the content push ratio corresponding to each of the various types of content within the second time period based on the output effect scores corresponding to the various types of content.

3. The method according to claim 1 or 2, characterized in that, The step of obtaining the output effect evaluation information of the first content among the content pushed within the first time period includes: Obtain the scores of each feature parameter of the first content in the content pushed within the first time period and the weights corresponding to each feature parameter; The output effect score corresponding to the first content is determined based on the scores of each feature parameter of the first content and the weights corresponding to each feature parameter.

4. The method according to claim 3, characterized in that, The first content includes video content, and the characteristic parameters of the video content include at least one of the following: first frame start-up status, playability status, frame rate status, stuttering status, CPU usage status, and memory usage status.

5. The method according to claim 4, characterized in that, The step of obtaining the score of each feature parameter of the first content in the content pushed within the first time period includes at least one of the following: Based on the difference between the first frame display time corresponding to the first video content within the first time period and the time when the terminal device detects the first operation triggering the playback of the first video content, a score is determined for the first frame playback status corresponding to the first video content. Based on whether the first video content was successfully played within the first time period, a score is obtained for the playability of the first video content. A score is determined based on the difference between the frame rate of the first video content and the refresh rate of the display screen of the terminal device during the first time period. Based on the number of times the first video content was interrupted within the first time period, a score for the interruption status of the first video content was obtained; A score for the CPU usage of the first video content is determined based on the ratio of the CPU resources occupied by the first video content during playback to the total CPU resources of the terminal device within the first time period. A score is determined based on the ratio of memory resources occupied by the first video content during playback to the total memory resources of the terminal device within the first time period.

6. The method according to claim 4 or 5, characterized in that, The step of obtaining the score of each feature parameter of the first content in the content pushed within the first time period includes: Obtain the scores of each feature parameter corresponding to multiple video contents within the first time period; The scores of each feature parameter of the video type content within the first time period are determined based on the scores of each feature parameter corresponding to the multiple video contents within the first time period.

7. The method according to any one of claims 3-6, characterized in that, The first content includes audio content, and the characteristic parameters of the audio content include at least one of the following: first frame start-up status, playability status, stuttering status, CPU usage status, and memory usage status.

8. The method according to any one of claims 3-7, characterized in that, The first content includes e-book type content, and the characteristic parameters of the e-book type content include at least one of the following: homepage display status, opening status, lag status, CPU usage status, and memory usage status.

9. The method according to any one of claims 3-8, characterized in that, The first content includes image-type content, and the characteristic parameters of the image-type content include at least one of image loading time, opening status, lag status, CPU usage, and memory usage.

10. The method according to claim 1, characterized in that, The output effect evaluation information of the first content includes the scores of each feature parameter corresponding to the first content; The method by which the cloud-side device determines the content push ratio of the first content within the second time period based on the output effect evaluation information includes: The cloud-side device obtains the weights of each feature parameter corresponding to the first content within the first time period; The cloud-side device determines the output effect score corresponding to the first content based on the scores of each feature parameter corresponding to the first content and the weights of each feature parameter corresponding to the first content. The cloud-side device determines the content push ratio of the first content within the second time period based on the output effect score corresponding to the first content.

11. The method according to any one of claims 1-10, characterized in that, The method further includes: When the first application of the terminal device is launched, the terminal device sends a first request to the cloud device corresponding to the first application. The first request is used to obtain evaluation strategy information, which includes each feature parameter of the first content and the weight corresponding to each feature parameter. The cloud-side device sends the evaluation strategy information to the edge-side device; The step of obtaining the output effect evaluation information of the first content among the content pushed within the first time period includes: Based on the evaluation strategy information, obtain the output effect evaluation information of the first content among the content pushed within the first time period.

12. The method according to any one of claims 1-10, characterized in that, The step of obtaining the output effect evaluation information of the first content among the content pushed within the first time period includes: The evaluation information of the first content among the content pushed within the first time period is obtained based on the evaluation strategy information stored in the terminal device. The evaluation strategy information includes each feature parameter of the first content and the weight corresponding to each feature parameter.

13. The method according to any one of claims 1-12, characterized in that, The method by which the cloud-side device determines the content push ratio of the first content within the second time period based on the output effect evaluation information includes: The cloud-side device determines the pre-push ratio of the first content within the second time period based on the recommended feature parameters; The cloud-side device determines the reduction ratio of the first content within the second time period based on the output effect evaluation information; The cloud-side device determines the content push ratio of the first content within the second time period based on the pre-push ratio and the reduction ratio.

14. The method according to claim 13, characterized in that, The recommended feature parameters include at least one of the following: user features corresponding to the terminal device, the popularity of the pre-push content stored in the cloud device, and the value of the pre-push content.

15. An information recommendation method, characterized in that, For a system including end-side devices and cloud-side devices, the method includes: The edge device acquires the output effect evaluation information of the first content among the content pushed within the first time period; The end-side device sends the output effect evaluation information to the cloud-side device; The cloud-side device determines the content push ratio of the first content corresponding to the second time period based on the output effect evaluation information. The content push ratio is used to characterize the proportion of the first content in the pushed content, and determines the content pushed to the end-side device based on the content push ratio.

16. An information recommendation system, characterized in that, include: The information recommendation system, comprising end-side devices and cloud-side devices, is used to execute the information recommendation method of claim 15.

17. An end-side device, characterized in that, include: One or more processors; One or more memories; the one or more memories storing one or more programs, which, when executed by the one or more processors, cause the end-side device to perform the information recommendation method according to any one of claims 1-14.

18. A readable storage medium, characterized in that, The readable storage medium stores instructions that, when executed on the end-side device, cause the end-side device to perform the information recommendation method according to any one of claims 1-14.

Images