Video recommendation system capable of improving real-time participation and reducing algorithm dependence

Abstract

The utility model discloses a video recommendation system capable of improving real-time participation and reducing algorithm dependence, and aims to improve real-time participation and reduce algorithm dependence so as to adapt to variable interests of users. The system is mainly characterized in that 1, recommendation content is dynamically adjusted in real time: the system dynamically adjusts the recommendation content according to real-time feedback information of a user in different time periods through a floating window recommendation mode, and ensures that the recommendation content always conforms to the current interest and mood of the user; the method overcomes the limitation that a traditional recommendation system can only depend on historical behavior data of the user for prediction. And 2, the algorithm complexity is reduced: as the real-time feedback information of the user is utilized to carry out recommendation adjustment and does not completely depend on a complex prediction algorithm, the dependence on a high-performance algorithm is reduced, the design and the implementation of a recommendation system are simplified, and the response speed and the flexibility of the system are improved.

Description

Technical Field

[0001] The utility model relates to the technical fields of video recommendation systems and algorithm optimization, and particularly relates to a video recommendation system that improves real-time participation and reduces algorithm dependence. Background Art

[0002] In current video recommendation systems, such as short video platforms like Douyin and Kuaishou, as well as medium and long video platforms like iQIYI and Tencent Video, the recommended content mainly relies on algorithmic autonomous recommendation. These algorithms usually process and analyze the historical behavior data of users, such as viewing records, click records, search records, like and comment records, etc., through machine learning models to predict the interests and preferences of users, thereby generating a list of recommended videos.

[0003] However, the existing recommendation systems have the following main problems:

[0004] 1. High dependence on algorithms: The existing recommendation systems highly rely on the accuracy and complexity of the recommendation algorithms. If the algorithms are not powerful enough, it is difficult to accurately deduce the preferences of users, and the relevance and accuracy of the recommended content will be greatly reduced. Especially in the case where users' interests change rapidly, it is difficult for the recommendation algorithms to capture the current needs of users in real time, resulting in the recommended content not meeting the actual interests of users.

[0005] 2. Difficulty in adapting to the changing interests of users: Users may have different interests and needs at different times. Traditional recommendation systems mainly rely on the historical behavior data of users for prediction, and it is difficult to capture and adapt to the current interests and moods of users in real time. This limitation leads to insufficient diversity and flexibility of the recommended content and poor user experience.

[0006] 3. High algorithm complexity and slow system response speed: In order to improve the accuracy of recommendations, existing systems usually need to use complex prediction algorithms and a large amount of computing resources. This not only increases the difficulty of system design and implementation, but also may lead to a slow system response speed, unable to adjust the recommended content in a timely manner, affecting the user experience.

[0007] 4. Lack of real-time feedback mechanism: Existing recommendation systems usually lack an effective real-time feedback mechanism. After users see the recommended content, if they find that it does not match their interests, they cannot immediately feedback to the system for adjustment. In this case, the system cannot optimize according to the real-time feedback information of users, and it is difficult to improve the accuracy and relevance of the recommended content.

[0008] Therefore, how to improve the accuracy and relevance of the recommended content has become the technical problem to be solved by the utility model. Content of the Utility Model

[0009] The technical problem solved by the present utility model is to provide a video recommendation system that improves real-time participation and reduces algorithm dependence to address the problem that the recommendation system in the above-mentioned background technology is difficult to capture the user's current moment needs in real time, aiming at the defects existing in the above-mentioned prior art.

[0010] To solve the above technical problems, the technical solutions adopted by the present utility model are as follows:

[0011] A video recommendation system that improves real-time participation and reduces algorithm dependence, including:

[0012] User device: The user device is used to display recommended content;

[0013] Application: Installed on the user device, used to execute the recommendation algorithm, and recommend at least one option to the user through the display area;

[0014] User feedback module: The user feedback module is used to receive the user's feedback information on the recommended content;

[0015] Recommendation adjustment module: The recommendation adjustment module communicates with the user feedback module and adjusts the parameters of the recommendation algorithm in real time based on, including but not limited to, the user's feedback information on the recommended content and each step of the user's operation in the application, so as to generate at least one new recommended option to ensure that the content of the recommended option is updated instantaneously according to the user's operation, making it more in line with the user's interests and needs during the current time period.

[0016] As a further solution of the present utility model, the video recommendation system includes, but is not limited to, short video sharing platforms such as Douyin and Kuaishou, and medium and long video platforms such as iQIYI and Tencent;

[0017] When the video recommendation system is used for short video sharing platforms such as Douyin and Kuaishou, short videos are recommended to the user through a floating window. When the user selects to accept a certain recommendation, the system will immediately give feedback and adjust the recommended content to be more in line with the user's current interests, thereby reducing the dependence on the initial algorithm.

[0018] When the video recommendation system is used for medium and long video platforms such as iQIYI and Tencent, the video recommendation system forms a recommendation floating window based on the user's past usage habits. When the user selects to accept a certain recommendation, the system will immediately give feedback and adjust the recommended content to be more in line with the user's current interests, thereby reducing the dependence on the initial algorithm.

[0019] As a further solution of the present utility model, at least one video option recommended by the floating window includes, but is not limited to, one or a combination of the following: the general direction of the video, the specific content of the video, the theme of the video, the category of the video, the tag of the video, the keyword of the video, or the user's current geographical location.

[0020] As a further solution of the present utility model, the feedback information received by the user feedback module includes, but is not limited to, the confirmation, rejection or rating of the recommended content by the user.

[0021] As a further solution of the present utility model, the recommendation algorithm module and the user feedback module are implemented by different independent application programs. The user manually inputs feedback information to the recommendation adjustment module, and the recommendation adjustment module is used to adjust the parameters of the recommendation algorithm according to the manually input feedback information.

[0022] As a further solution of the present utility model, the recommendation algorithm and the user feedback module are integrated on the cloud server. The user device is only used to display the recommended content and submit feedback information. The feedback information is processed by the cloud server to adjust the recommendation algorithm, so as to realize the update of the recommended content.

[0023] Compared with the prior art, the beneficial effects of the present utility model are as follows:

[0024] 1. Adapt to the changing interests of users: The present utility model takes into account that users may have different interests at different times. Through the floating window recommendation method, it can dynamically adjust the recommended content according to the real-time feedback information of users, ensuring that the recommended content always conforms to the current interests and moods of users. This method overcomes the limitation that traditional recommendation systems can only rely on users' historical behavior data for prediction.

[0025] 2. Reduce the algorithm complexity: Since the present utility model adjusts the recommendation through the real-time feedback information of users and no longer completely relies on complex prediction algorithms, this not only reduces the dependence on high-performance algorithms, but also simplifies the design and implementation of the recommendation system, improving the response speed and flexibility of the system.

[0026] 3. Applicability to multiple platforms: The video recommendation system of the present utility model is not only applicable to short video platforms such as Douyin and Kuaishou, but also applicable to medium and long video platforms such as iQIYI and Tencent Video. Through the floating window recommendation method, whether it is short videos or medium and long videos, the system can capture the feedback information of users in real time and adjust the recommendation, so that it can be effectively applied in different types of video platforms.

[0027] 4. Flexible recommendation options: At least one video option in the floating window recommendation can include various information such as the general direction of the video, specific content, theme, category, tag, keyword or the current geographical location of the user, providing flexible and diverse recommended content to meet the diverse needs of users.

[0028] The additional aspects and advantages of the present utility model will be partly given in the following description, partly become obvious from the following description, or be understood through the practice of the present utility model. Description of the Drawings

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0030] Figure 1 It is a schematic diagram of the cooperation of the video recommendation system module of the present invention.

[0031] Figure 2 It is a schematic diagram of the cooperation of the video recommendation method module of the present invention. Specific embodiments

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0033] Please refer to Figure 1 -2. In the embodiments of the present invention, a video recommendation system that improves real-time participation and reduces algorithm dependence includes:

[0034] User device: The user device is used to display recommended content; the user device includes a display area.

[0035] Application program: Installed on the user device, used to execute the recommendation algorithm, and recommend at least one option to the user through the display area, for example, in the form of a semi-transparent floating window or a fixed window.

[0036] User feedback module: The user feedback module is used to receive the feedback information of the user on the recommended content.

[0037] Recommendation adjustment module: The recommendation adjustment module communicates with the user feedback module, and based on, including but not limited to, the feedback information of the user on the recommended content and each operation of the user in the application program, adjusts the parameters of the recommendation algorithm in real time to generate at least one new recommended option to ensure that the content of the recommended option is updated immediately according to the user's operation, so as to better meet the interests and needs of the user in the current time period.

[0038] As a further solution of the present invention, the video recommendation system includes, but is not limited to, short video sharing platforms such as Douyin and Kuaishou, and medium and long video platforms such as iQIYI and Tencent.

[0039] When the video recommendation system is used in short video sharing platforms such as Douyin and Kuaishou, short videos are recommended to users through floating windows. When the user selects to accept a certain recommendation, the system will immediately give feedback and adjust the recommended content to better match the user's current interests, thereby reducing the dependence on the initial algorithm.

[0040] When the video recommendation system is used in medium and long video platforms such as iQIYI and Tencent, the video recommendation system forms a recommendation floating window based on the user's past usage habits. When the user selects to accept a certain recommendation, the system will immediately give feedback and adjust the recommended content to better match the user's current interests, thereby reducing the dependence on the initial algorithm.

[0041] As a further solution of the present utility model, at least one video option recommended by the floating window includes, but is not limited to, one or a combination of the following: the general direction of the video, the specific content of the video, the theme of the video, the category of the video, the tag of the video, the keyword of the video, or the user's current geographical location.

[0042] As a further solution of the present utility model, the feedback information received by the user feedback module includes, but is not limited to, the user's confirmation, rejection, or rating of the recommended content.

[0043] As a further solution of the present utility model, the recommendation algorithm module and the user feedback module are implemented by different independent application programs. The user manually inputs feedback information to the recommendation adjustment module, and the recommendation adjustment module is used to adjust the parameters of the recommendation algorithm according to the manually input feedback information.

[0044] As a further solution of the present utility model, the recommendation algorithm and the user feedback module are integrated on the cloud server. The user device is only used to display the recommended content and submit feedback information. The feedback information is processed by the cloud server to adjust the recommendation algorithm, thereby realizing the update of the recommended content.

[0045] A video recommendation method for reducing algorithm dependence and improving recommendation accuracy includes the following steps:

[0046] a. Display video recommendation content on the user device;

[0047] b. Execute the recommendation algorithm, which can run on, including but not limited to, the user device, the cloud server, and the remote server;

[0048] c. Recommend at least one video option to the user through a floating window;

[0049] d. Receive the user's feedback information on the recommended content;

[0050] e. Based on the user feedback information, adjust the parameters of the recommendation algorithm in real time to update and optimize the recommended content.

[0051] As a further solution of the present utility model, the video recommendation system is applicable to, including but not limited to, short video sharing platforms and medium- and long-video platforms, where:

[0052] In the short video sharing platform, short videos are recommended to users through floating windows; when the user selects to accept a certain recommendation, the system immediately gives feedback and adjusts the recommended content to better match the user's current interests, thereby reducing the dependence on the initial algorithm.

[0053] In the medium- and long-video platform, the video recommendation system forms a recommendation floating window based on the user's past usage habits, that is, in the medium- and long-video platform, video content is recommended to users through floating windows; when the user selects to accept a certain recommendation, the system immediately gives feedback and adjusts the recommended content to better match the user's current interests, thereby reducing the dependence on the initial algorithm.

[0054] As a further solution of the present utility model, the past usage habits include but are not limited to the video the user last opened, the videos the user clicked on, the viewing duration of the user, the user's search history, and the user's like or comment records on the video content.

[0055] As a further solution of the present utility model, the execution of the recommendation algorithm includes the following steps:

[0056] a. Collect and analyze the user's historical behavior data, including but not limited to viewing records, click records, search records, like and comment records;

[0057] b. Use a machine learning model to process and analyze the collected data to predict the user's interests and preferences;

[0058] c. Generate a recommended video list and display it to the user through a floating window;

[0059] d. Dynamically adjust the parameters of the recommendation algorithm based on the user's real-time feedback information to improve the accuracy and relevance of the recommended content.

[0060] Embodiment 1:

[0061] This embodiment provides a video recommendation system that improves real-time participation and reduces algorithm dependence. The system includes a user device, an application program, a user feedback module, and a recommendation adjustment module.

[0062] User device: The user device can be a terminal device such as a smartphone, a tablet computer, or a computer, mainly used to display video recommendation content. The user watches videos through the user device and receives recommended video options through the application program.

[0063] Application: The application is installed on the user device and is responsible for executing the recommendation algorithm and recommending at least one video option to the user through a floating window. The application utilizes the user's historical behavior data and machine learning models to generate a preliminary list of recommended videos, which is then presented to the user in the form of a floating window.

[0064] User feedback module: The user feedback module is integrated into the application and is used to receive feedback information from the user regarding the recommended content. The feedback information can include the user's confirmation, rejection, or rating of the recommended videos, etc. The user completes the feedback by clicking on the options in the floating window or manually entering the feedback information.

[0065] Recommendation adjustment module: The recommendation adjustment module communicates with the user feedback module and, upon receiving the feedback information, adjusts the parameters of the recommendation algorithm in real time. The recommendation adjustment module optimizes the recommendation algorithm based on the user's feedback information, enabling the system to dynamically update and optimize the recommended content according to the user's current interests. The recommendation adjustment module can run on a cloud server and communicate with the user device via the network to ensure real-time processing and adjustment.

[0066] For example, the application scenarios of short video platforms (such as Douyin and Kuaishou): On short video platforms, the user device displays the recommended short videos. The application recommends at least one short video option to the user through a floating window, and the user can immediately provide feedback on the recommended video. When the user selects a certain recommended video, the system will immediately adjust the parameters of the recommendation algorithm based on the user's feedback and update the recommended content in real time to ensure that the next recommendation better matches the user's current interests.

[0067] Another example is the application scenarios of medium and long video platforms (such as iQiyi and Tencent Video): On medium and long video platforms, the user device displays the recommended medium and long videos. The application recommends video content through a floating window based on the user's past usage habits. The user can provide feedback on the recommended video through the floating window, and the system will adjust the parameters of the recommendation algorithm in real time based on the user's feedback information to optimize the recommended content. In this way, whether it is a short video platform or a medium and long video platform, the system can capture the user's interest changes in real time and provide more accurate recommended content.

[0068] Through the above embodiments, those of ordinary skill in the art can accurately implement the present utility model, thereby significantly improving the real-time participation and recommendation accuracy of the video recommendation system, reducing the dependence on complex algorithms, simplifying the system design, and improving the response speed.

[0069] Embodiment 2:

[0070] The specific algorithm implementation of a video recommendation system that improves real-time participation and reduces algorithm dependence is as follows:

[0071] Data collection: The system first collects the user's historical behavior data, including but not limited to viewing records, click records, search records, like and comment records, etc. This data is continuously collected through the application and stored on the cloud server for use by the recommendation algorithm.

[0072] Initial recommendation generation: The system uses the collaborative filtering algorithm to generate initial recommendations. The collaborative filtering algorithm is based on user similarity and item similarity, analyzing which videos are commonly liked by similar users, thus generating a preliminary list of recommended videos.

[0073] For example, for a certain user, the system will find other users with similar behaviors, view the videos liked by these users, and then recommend these videos to that user.

[0074] Float window recommendation: The application runs on the user's device and recommends at least one video option to the user through a float window. The content of the float window recommendation can include the general direction of the video, specific content, theme, category, tag, keyword, or the user's current geographical location, etc.

[0075] This way of float window recommendation provides a quick feedback channel for users, enabling the system to capture the user's immediate interests and enhancing the system's response speed and flexibility.

[0076] User feedback collection: Users provide feedback on the recommended content through the float window. The feedback forms can be confirmation, rejection, or rating, etc. The user feedback module receives the user's feedback information and sends this information to the recommendation adjustment module.

[0077] This real-time feedback mechanism enables the system to dynamically adjust the recommended content according to the user's immediate feedback, improving the accuracy and relevance of the recommendation and reducing the dependence on the initial algorithm.

[0078] Real-time adjustment algorithm: The recommendation adjustment module adjusts the parameters of the recommendation algorithm in real time according to the user's feedback information. The adjustment process includes retraining or updating the parameters of the collaborative filtering model. For example, if a user rejects a recommended video, the system will reduce the weight of videos with similar features to that video in the recommendation list; if the user confirms or gives a high rating, the system will increase the recommendation weight of similar videos.

[0079] In addition, the system can also adopt reinforcement learning algorithms to dynamically adjust the recommendation strategy according to the user's feedback information, enabling the recommendation system to continuously self-optimize and improve the recommendation accuracy. In this way, the system can adapt to the changing interests of users at different times, ensuring that the recommended content always conforms to the user's current interests and mood, and overcoming the limitation that traditional recommendation systems can only rely on historical behavior data for prediction.

[0080] Recommendation content update: The adjusted recommendation algorithm generates a new list of recommended videos and displays the updated recommendation content to the user again through the floating window of the application. This loop process ensures that the system can dynamically adapt to the changing interests of users and provide more accurate recommendation content.

[0081] Its application scenarios and effects, such as short video platforms (e.g., Douyin, Kuaishou):

[0082] Short videos are recommended through the floating window, and the system immediately provides feedback for adjustment after the user makes a selection. Through the real-time feedback mechanism, the system can dynamically update the recommended content according to the user's immediate interests, improve the accuracy and relevance of the recommendations, and reduce the dependence on the initial algorithm.

[0083] Also, when its application scenario is in medium and long video platforms (e.g., iQiyi, Tencent Video): Medium and long video content is recommended through the floating window, and the system immediately provides feedback for adjustment after the user makes a selection. The system dynamically adjusts the recommended content based on the user's real-time feedback information to ensure that the recommended content always matches the user's current interests and mood, overcoming the limitation that traditional recommendation systems can only rely on historical behavior data for prediction.

[0084] Reduce algorithm complexity: By adjusting the recommendation algorithm parameters based on the user's real-time feedback information, the dependence on complex prediction algorithms is reduced, the design and implementation of the recommendation system are simplified, and the response speed and flexibility of the system are improved.

[0085] Multi-platform applicability: Whether it is a short video or a medium and long video platform, the system can capture the user's feedback information in real time and make recommendation adjustments, so it can be effectively applied in different types of video platforms.

[0086] Through the above embodiments, those of ordinary skill in the art can accurately implement the specific algorithm implementation of the present utility model, thereby significantly improving the real-time participation and recommendation accuracy of the video recommendation system, reducing the dependence on complex algorithms, simplifying the system design, and improving the response speed.

[0087] Embodiment 3:

[0088] This embodiment provides an implementation for presenting multiple reference scenarios, which includes:

[0089] Data collection: The system first collects the user's historical behavior data, including but not limited to viewing records, click records, search records, like and comment records, etc. These data are continuously collected through the application and stored on the cloud server for use by the recommendation algorithm.

[0090] Initial Recommendation Generation: The system uses collaborative filtering algorithms to generate initial recommendations. Collaborative filtering algorithms are based on user similarity and item similarity, analyzing which videos are commonly liked by similar users, thus generating an initial list of recommended videos. For example, for a certain user, the system will find other users with similar behaviors, view the videos liked by these users, and then recommend these videos to that user.

[0091] Floating Window Recommendation: The application runs on the user's device and recommends at least one video option to the user through a floating window. The content of the floating window recommendation can include the general direction of the video, specific content, theme, category, tag, keyword, or the user's current geographical location, etc. This way of floating window recommendation provides a quick feedback channel for users, enabling the system to capture the user's immediate interests and enhancing the system's response speed and flexibility.

[0092] User Feedback Collection: Users provide feedback on the recommended content through the floating window. The feedback forms can be confirmation, rejection, or rating, etc. The user feedback module receives the feedback information from users and sends this information to the recommendation adjustment module. The real-time feedback mechanism enables the system to dynamically adjust the recommended content according to the user's immediate feedback, improving the accuracy and relevance of the recommendation and reducing the dependence on the initial algorithm.

[0093] Application Scenario 1: Implemented by independent applications.

[0094] The recommendation algorithm module and the user feedback module are implemented by different independent applications:

[0095] In this solution, two independent applications run on the user's device: one for executing the recommendation algorithm and the other for collecting user feedback.

[0096] After seeing the recommended content, the user can manually input feedback information into the user feedback module.

[0097] The user feedback module sends the feedback information to the recommendation adjustment module, and the recommendation adjustment module adjusts the parameters of the recommendation algorithm according to the manually input feedback information.

[0098] For example, the user can input "like" or "dislike" a certain recommended video through the feedback application, and the system adjusts the recommendation algorithm based on this feedback information to generate a new recommended list.

[0099] This implementation method of independent applications simplifies the process of collecting and processing user feedback, ensuring that the system can quickly adjust the recommended content according to the user's immediate feedback.

[0100] Application Scenario 2: Cloud Integration.

[0101] The recommendation algorithm and the user feedback module are integrated on the cloud server:

[0102] In this solution, both the recommendation algorithm and the user feedback module are integrated on the cloud server, and the user device is only used to display the recommended content and submit feedback information.

[0103] The user submits feedback information through the user device, and these feedback information are sent to the cloud server for processing.

[0104] The cloud server adjusts the parameters of the recommendation algorithm according to the feedback information, updates the recommended content, and sends the new recommendation list back to the user device.

[0105] For example, when the user is watching a recommended video, they can submit feedback information through the floating window on the user device, and these information are uploaded to the cloud server in real time for processing. The cloud server dynamically adjusts the recommendation algorithm according to these feedback information to ensure that the recommended content is more accurate and relevant.

[0106] This way of cloud integration not only improves the processing efficiency and response speed of the system, but also reduces the computing burden on the user device, ensuring that the system can operate efficiently on different platforms.

[0107] Through the above embodiments, those of ordinary skill in the art can accurately implement the specific algorithm implementation of the present utility model, thereby significantly improving the real-time participation and recommendation accuracy of the video recommendation system, reducing the dependence on complex algorithms, simplifying the system design, and improving the response speed. This method is applicable to a variety of video platforms, including short video and medium and long video platforms, and can effectively improve the user experience.

[0108] Embodiment 4:

[0109] This embodiment provides a video recommendation system and method for entering a specific interface after the user clicks on a recommendation.

[0110] Data collection: The system first collects the user's historical behavior data, including but not limited to viewing records, click records, search records, like and comment records, etc. These data are continuously collected through the application program and stored on the cloud server for use by the recommendation algorithm.

[0111] Initial recommendation generation: The system uses the collaborative filtering algorithm to generate an initial recommendation. The collaborative filtering algorithm is based on user similarity and item similarity, and analyzes which videos are commonly liked by similar users, thereby generating an initial recommended video list.

[0112] Floating window recommendation: The application program runs on the user device and recommends at least one video option to the user through the floating window. The content of the floating window recommendation can include the general direction of the video, specific content, theme, category, tag, keyword, or the user's current geographical location, etc.

[0113] User feedback collection: Users provide feedback on the recommended content through the floating window. The feedback forms can include confirmation, rejection, rating, etc. The user feedback module receives the feedback information from users and sends this information to the recommendation adjustment module.

[0114] Real-time adjustment algorithm: The recommendation adjustment module adjusts the parameters of the recommendation algorithm in real-time according to the feedback information from users. For example, if a user rejects a recommended video, the system will reduce the weight of videos with similar features to that video in the recommendation list; if a user confirms or gives a high rating, the system will increase the recommendation weight of similar videos.

[0115] Update of recommended content: The adjusted recommendation algorithm generates a new list of recommended videos and displays the updated recommended content to users again through the floating window of the application.

[0116] Jump after click: When a user selects a certain recommended video, the system will direct the user to a specific interface according to the type of the video or the platform:

[0117] Short video platforms (such as Douyin, Kuaishou): After the user clicks on the recommended video, the system directly jumps to the playing page of the video on the short video platform, and the user can immediately watch the recommended video.

[0118] If the recommended video is related to a product, it can be jumped to the Douyin Mall or the relevant product details page, where users can further understand the product information or make a purchase.

[0119] Medium and long video platforms (such as iQiyi, Tencent Video): After the user clicks on the recommended video, the system directly jumps to the playing page of the video on the medium and long video platform, and the user can immediately watch the recommended video.

[0120] If the recommended content is a drama or a movie, the system can jump to the details page of the drama or the movie, where users can view the detailed information, relevant reviews, or select to watch other episodes.

[0121] Association with e-commerce platforms (such as Douyin Mall): If the video recommendation system is integrated with an e-commerce platform (such as Douyin Mall), when the user clicks on the recommended video, it can directly jump to the details page of the relevant product.

[0122] On the details page, users can view product information, prices, user reviews, etc., and there is a purchase button for users to place an order directly.

[0123] This design of jumping after click not only improves the convenience of user use but also effectively enhances the actual conversion rate and user experience of the video recommendation system. By adjusting the recommendation algorithm in real-time and guiding users to enter specific interfaces, the system can better meet the immediate needs of users and improve the accuracy and relevance of the recommended content.

[0124] Through the above embodiments, those of ordinary skill in the art can accurately implement the present utility model, thereby significantly improving the real-time participation and recommendation accuracy of the video recommendation system, reducing the dependence on complex algorithms, simplifying the system design, and improving the response speed. At the same time, by entering a specific interface after clicking on the recommendation, the user experience and the conversion effect of the system are enhanced.

[0125] For those skilled in the art, it is obvious that the present utility model is not limited to the details of the above-described exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present utility model. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present utility model is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present utility model.

Claims

1. A video recommendation system that improves real-time participation and reduces algorithm dependence, characterized in that Including: User device: The user device is used to display recommended content; Application: Installed on the user device, used to execute a recommendation algorithm, and recommend at least one option to the user through a display area; User feedback module: The user feedback module is used to receive feedback information from the user on the recommended content; Recommendation adjustment module: The recommendation adjustment module communicates with the user feedback module and adjusts the parameters of the recommendation algorithm in real time based on, including but not limited to, the feedback information of the user on the recommended content and each operation of the user in the application, so as to generate at least one new recommended option to ensure that the content of the recommended option is updated immediately according to the user's operation, so as to better meet the interests and needs of the user in the current time period.

2. The video recommendation system for improving real-time participation and reducing algorithm dependence according to claim 1, wherein The video recommendation system includes, but is not limited to, short video sharing platforms such as Douyin and Kuaishou, and medium and long video platforms such as iQIYI and Tencent; When the video recommendation system is used for short video sharing platforms such as Douyin and Kuaishou, short videos are recommended to the user through a floating window; when the user selects to accept a certain recommendation, the system will immediately give feedback and adjust the recommended content to better match the user's current interests, thereby reducing the dependence on the initial algorithm; When the video recommendation system is used for medium and long video platforms such as iQIYI and Tencent, the video recommendation system forms a recommendation floating window according to the user's past usage habits. When the user selects to accept a certain recommendation, the system will immediately give feedback and adjust the recommended content to better match the user's current interests, thereby reducing the dependence on the initial algorithm.

3. A video recommendation system for improving real-time participation and reducing algorithm dependence according to claim 1, characterized in that, The at least one video option recommended by the floating window includes, but is not limited to, one or a combination of the following: the general direction of the video, the specific content of the video, the theme of the video, the category of the video, the tag of the video, the keyword of the video, or the user's current geographical location.

4. A video recommendation system for improving real-time participation and reducing algorithm dependence according to claim 1, characterized in that, The feedback information received by the user feedback module includes, but is not limited to, the user's confirmation, rejection, or rating of the recommended content.

5. A video recommendation system that improves real-time participation and reduces algorithm dependence according to claim 1, characterized in that, The recommendation algorithm module and the user feedback module are implemented by different independent applications. The user manually inputs feedback information to the recommendation adjustment module, and the recommendation adjustment module is used to adjust the parameters of the recommendation algorithm according to the manually input feedback information.

6. A video recommendation system for improving real-time participation and reducing algorithm dependence according to claim 1, characterized in that, The recommendation algorithm and the user feedback module are integrated on the cloud server. The user device is only used to display the recommended content and submit feedback information. The feedback information is processed by the cloud server to adjust the recommendation algorithm, so as to realize the update of the recommended content.

Images