Video recommendation method and device, equipment and storage medium

Abstract

The invention relates to a video recommendation method and device, equipment and a storage medium, in the application, an instruction triggered by a current user for accessing a target scene is received, and if the current user is an inactive user in the target scene, a candidate scene corresponding to the target scene is determined; wherein the candidate scene is a scene similar to a target scene mode; if the current user is an active user in the candidate scene, recommending a video to the current user through a recommendation model of the candidate scene; and if the current user is an inactive user in the candidate scene, recommending a video to the current user through the recommendation model of the target scene. Visibly, when the user enters the scene with less access, the video can be recommended through the model of the candidate scene with the similar mode, due to the fact that the modes of the two scenes are similar, when the video is recommended by switching the model of the candidate scene, the user does not have obvious perception, in addition, the video recommendation effect can be improved through the model of the candidate scene, and the user experience is improved. And user experience is improved.

Description

Technical Field

[0001] This application relates to the field of video recommendation, and particularly to a video recommendation method, apparatus, device, and storage medium. Background Art

[0002] Currently, in the short video recommendation function of applications, there are multiple scenarios. When a user accesses a certain scenario, the model of that scenario is required to recommend videos for the user. When the user accesses scenarios, they may access multiple scenarios. If the user is a deep-access user in a certain scenario, the model of that scenario will collect more user behavior data. Therefore, the model is well-trained and the recommendation accuracy is relatively high. However, if the user enters a scenario that is rarely visited or has never been visited before, due to the small amount of user behavior data collected by the new scenario, the model is not well-trained, resulting in low accuracy of the videos recommended to the user and a reduced user experience.

[0003] Therefore, how to improve the accuracy of video recommendation and enhance the user experience is a problem that needs to be solved by those skilled in the art. Summary of the Invention

[0004] This application provides a video recommendation method, apparatus, device, and storage medium to improve the accuracy of video recommendation and the user experience.

[0005] In a first aspect, this application provides a video recommendation method, including:

[0006] Receiving an access instruction triggered by a current user; wherein, the access instruction is an instruction for the current user to access a target scenario;

[0007] If the current user is an inactive user in the target scenario, determining a candidate scenario corresponding to the target scenario; wherein, the candidate scenario is a scenario with a similar pattern to the target scenario;

[0008] If the current user is an active user in the candidate scenario, recommending videos for the current user through the recommendation model of the candidate scenario.

[0009] Optionally, before receiving the access instruction triggered by the current user, it further includes:

[0010] Determining candidate scenarios corresponding to each scenario.

[0011] Optionally, determining candidate scenarios with a similar pattern to each scenario includes:

[0012] Determining screening scenarios with a similar pattern to each scenario;

[0013] If the number of screening scenarios is one, using the screening scenario as the candidate scenario;

[0014] If the number of screening scenarios is greater than one, determine the activity level of the user in each screening scenario, and use the screening scenario with the highest activity level as the candidate scenario.

[0015] Optionally, after receiving the access instruction triggered by the current user, it includes:

[0016] If the current user is an active user in the target scenario, recommend videos for the current user through the recommendation model of the target scenario;

[0017] If the current user is an inactive user in the candidate scenario, recommend videos for the current user through the recommendation model of the target scenario.

[0018] Optionally, before receiving the access instruction triggered by the current user, it further includes:

[0019] Determine the user types of each user in different scenarios; wherein, the user types include: active users and inactive users.

[0020] Optionally, the determining the user types of each user in different scenarios includes:

[0021] Determine the number of accesses of each user in different scenarios; wherein, the number of accesses is: the number of times each user accesses different scenarios within a predetermined time period;

[0022] Determine the user types of users in different scenarios according to the number of accesses of each user in different scenarios and a predetermined threshold.

[0023] Optionally, after determining the user types of each user in different scenarios, it further includes:

[0024] Determine the target users corresponding to each scenario according to the user types of each user in different scenarios; wherein, the target users are: users whose user type in each scenario is inactive and whose user type in the corresponding candidate scenario is active;

[0025] Correspondingly, if the current user is an active user in the candidate scenario, recommending videos for the current user through the recommendation model of the candidate scenario includes:

[0026] If the current user is the target user of the target scenario, determine that the current user is an active user in the candidate scenario, and recommend videos for the current user through the recommendation model of the candidate scenario.

[0027] In a second aspect, the present application provides a video recommendation device, including:

[0028] A receiving module, configured to receive an access instruction triggered by a current user; wherein, the access instruction is an instruction for the current user to access a target scenario;

[0029] A first determination module, configured to determine a candidate scenario corresponding to the target scenario when the current user is an inactive user in the target scenario; wherein, the candidate scenario is a scenario with a similar pattern to the target scenario;

[0030] A first recommendation module, configured to recommend a video for the current user through a recommendation model of the candidate scenario when the current user is an active user in the candidate scenario.

[0031] In a third aspect, the present application provides an electronic device, including:

[0032] A processor, a memory, and a computer program stored on the memory and executable on the processor, and the processor executes the steps of the video recommendation method of the present application through the computer program.

[0033] In a fourth aspect, the present application further provides a computer storage medium, where the computer storage medium stores computer-executable instructions, and the computer-executable instructions are used to execute the steps of the video recommendation method of the present application.

[0034] The above technical solutions provided by the embodiments of the present application have the following advantages compared with the prior art: The present application discloses a video recommendation method, device, device, and storage medium. In the present application, an instruction for a current user to access a target scenario is received. If the current user is an inactive user in the target scenario, a candidate scenario corresponding to the target scenario is determined; wherein, the candidate scenario is a scenario with a similar pattern to the target scenario; if the current user is an active user in the candidate scenario, a video is recommended for the current user through the recommendation model of the candidate scenario; if the current user is an inactive user in the candidate scenario, a video is recommended for the current user through the recommendation model of the target scenario. It can be seen that when the user enters a scenario with less access, the present application can recommend videos through the model of the candidate scenario with a similar pattern. Since the patterns of these two scenarios are similar, when switching to the model of the candidate scenario to recommend videos, the user will not have an obvious perception. Moreover, this method can improve the video recommendation effect through the model of the candidate scenario and improve the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present invention and used together with the specification to explain the principles of the present invention.

[0036] 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, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0037] One or more embodiments are exemplarily illustrated by the pictures in the corresponding drawings. These exemplary illustrations do not limit the embodiments. Elements with the same reference numerals in the drawings are represented as similar elements. Unless otherwise stated, the drawings in the figures do not constitute a proportional limitation.

[0038] Figure 1 Schematic diagram of a video recommendation method provided by an embodiment of the present application;

[0039] Figure 2 Schematic diagram of another video recommendation method provided by an embodiment of the present application;

[0040] Figure 3 Schematic diagram of another video recommendation method provided by an embodiment of the present application;

[0041] Figure 4 Schematic diagram of another video recommendation method provided by an embodiment of the present application;

[0042] Figure 5 Schematic diagram of a recommendation result generation process provided by an embodiment of the present application;

[0043] Figure 6 Schematic diagram of the overall framework of a recommendation result generation process provided by an embodiment of the present application;

[0044] Figure 7 Schematic diagram of the structure of a video recommendation device provided by an embodiment of the present application;

[0045] Figure 8 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0047] The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. To simplify the disclosure of the present invention, components and settings of specific examples are described below. Of course, they are only examples and are not intended to limit the present invention. In addition, the present invention may repeat reference numerals and / or letters in different examples. Such repetition is for the purpose of simplicity and clarity and does not itself indicate the relationship between various embodiments and / or settings discussed.

[0048] Currently, since each scenario differs in terms of entry, user group, usage habits, etc., it is not possible to simply fuse multiple scenario samples together for training. Therefore, each scenario has a corresponding recommendation model, and each recommendation model is trained through user behavior data such as when users access that scenario, in order to accurately recommend videos for users. If a user enters a new scenario that they usually visit less frequently, due to less user behavior data being collected for the new scenario and insufficient model training, the accuracy of the videos recommended to the user will be low, reducing the user experience.

[0049] Therefore, in this application, a cross-scenario model switching scheme is proposed. This scheme mainly targets the situation where a user accesses multiple scenarios. If the user accesses an inactive scenario, candidate scenarios for the inactive scenario are determined, and videos are recommended to the user through the recommendation model of the candidate scenario, so as to improve the accuracy of video recommendation and the user experience.

[0050] See Figure 1 , which is a schematic flowchart of a video recommendation method provided by an embodiment of this application. The method specifically includes the following steps:

[0051] S101. Receive an access instruction triggered by the current user; wherein, the access instruction is an instruction for the current user to access a target scenario;

[0052] In this application, when a user accesses a certain scenario, an access instruction to access that scenario is triggered. This application obtains this access instruction in order to accurately recommend videos for the user. This application refers to the scenario currently accessed by the user as the target scenario. This scenario may include: an immersive playback scenario, a Feed stream scenario, etc., and is not specifically limited herein.

[0053] S102. If the current user is an inactive user in the target scenario, determine candidate scenarios corresponding to the target scenario; the candidate scenarios are scenarios with a similar pattern to the target scenario;

[0054] In this application, users can be divided into different user types in different scenarios according to their different levels of activity in different scenarios. The user types may include: active users and inactive users. If a user is highly active in a certain scenario, the user can be determined to be an active user in the scenario. If a user is less active in a certain scenario, the user can be determined to be an inactive user in the scenario.

[0055] After this application obtains the target scene access instruction triggered by the current user, it can determine the user type of the current user in the target scene. When this application determines the user type, the user type of the current user in the target scene can be determined in the following two ways: Method 1: Determine the user type of the current user in different scenes in advance based on the activity level information of the current user in different scenes; therefore, after receiving the access instruction triggered by the current user, the user type of the current user in the target scene can be directly determined; Method 2: After receiving the access instruction triggered by the current user, obtain the activity level information of the current user in the target scene in real time, and determine the user type of the current user in the target scene in real time based on the activity level information.

[0056] Moreover, the activity level information in this application is to identify the user's activity level in different scenarios through the user's historical behavior data, which can be the number of visits, visit time, etc. of the user to a certain scenario, and is not specifically limited here. For example: if the number of times user A visits scenario A exceeds a predetermined threshold within a period of time, then the user type of user A in scenario A can be determined to be an active user; if the number of times user A visits scenario A within a period of time is less than the predetermined threshold, then the user type of user A in scenario A can be determined to be an inactive user.

[0057] When the present application determines that the current user is an inactive user in the target scene, since the current user is less active in the target scene, if the recommendation model of the target scene is used to recommend videos to the current user, the accuracy of the recommended videos for the user will be low, which will reduce the user experience. Therefore, in the present application, the candidate scene of the target scene can be determined, and the candidate scene is a scene similar to the target scene mode, for example: the candidate scene of the immersive playback scene must also be in the immersive mode, and the candidate scene of the feed stream scene must also be in the feed stream mode.

[0058] In the present application, the candidate scenes of each scene can be determined by manual designation or automatic selection. The candidate scenes are selected by: scenes with similar patterns that are different from the original scene but have similar patterns.

[0059] S103: If the current user is an active user in the candidate scene, recommend videos to the current user through a recommendation model of the candidate scene.

[0060] In this application, after determining the candidate scenarios corresponding to the target scenario, the user type of the current user in the candidate scenario can be determined, and which recommendation model to use for video recommendation can be determined according to the user type of the current user in the candidate scenario. For example, if the current user is an inactive user in the candidate scenario, it means that the user's activity level in this candidate scenario is also very low. If the recommendation model for this candidate scenario is switched to recommend videos to the user, the recommendation accuracy will still be very low and the user experience cannot be improved. If the current user is an active user in the candidate scenario, it means that the user's activity level in this candidate scenario is relatively high. At this time, the recommendation model for this candidate scenario can be switched to recommend videos to the user, thereby improving the user experience.

[0061] This application can also provide a setting interface for the user. Through this setting interface, the user can set whether to start the recommendation function described in this application. If the user starts the recommendation function described in this application through this setting interface, after detecting the access instruction sent by the user, it is determined whether to switch the recommendation model of the candidate scenario. If the user does not start the recommendation function described in this application through this setting interface, after receiving the access instruction sent by the user, the recommendation model of the target scenario is directly used to recommend videos to the user, and there is no need to switch the recommendation model.

[0062] It should be noted that the user may access multiple scenarios, but the user's activity levels in different scenarios are different. Some scenarios are accessed less by the user, and some scenarios are accessed more by the user. For the scenarios that the user accesses more, the user's preferences can be better collected, so the recommendation effect of the recommendation model is better. On the contrary, for the scenarios that the user accesses less, the recommendation effect is worse. Therefore, this application can identify the user's activity levels in different scenarios through the user's historical behavior data. When the user accesses a scenario with a low activity level, it automatically switches to a candidate scenario with a higher activity level, and recommends videos to the user through the recommendation model of the candidate scenario. Even if there is a large difference between the sorting results of the recommendation model after switching and the original model results, the user will not have an obvious perception, because the user accesses the original scenario less and does not have a deep perception experience. Switching the model can improve the recommendation results of such users and optimize the user experience.

[0063] In summary, when the user enters a scenario with less access in this application, videos can be recommended through the model of a candidate scenario with similar patterns. Since the patterns of these two scenarios are similar, when switching the model of the candidate scenario to recommend videos, the user will not have an obvious perception. Moreover, this method can improve the video recommendation effect through the model of the candidate scenario and improve the user experience.

[0064] See Figure 2 , which is another schematic flowchart of the video recommendation method provided by the embodiment of this application. The method specifically includes the following steps:

[0065] S201, determining candidate scenes corresponding to each scene;

[0066] S202, receiving an access instruction triggered by the current user; wherein the access instruction is an instruction for the current user to access a target scene;

[0067] S203: If the current user is an inactive user in the target scene, determine a candidate scene corresponding to the target scene; the candidate scene is a scene with a similar pattern to the target scene;

[0068] S204: If the current user is an active user in the candidate scene, recommend videos to the current user through a recommendation model of the candidate scene.

[0069] In this embodiment, the candidate scenes corresponding to each scene can be determined in advance, so that after receiving the user's access instruction to access the target scene, the candidate scenes can be directly determined, thereby improving the speed of determining the candidate scenes; it should be noted that different users have different access habits, so the candidate scenes of different users in different scenes are also different.

[0070] In another embodiment of the present application, the process of determining candidate scenes similar to each scene pattern specifically includes: determining a screening scene similar to the pattern of each scene; if the number of screening scenes is one, taking the screening scene as a candidate scene; if the number of screening scenes is greater than one, determining the user's activity level in each screening scene, and taking the screening scene with the highest activity level as a candidate scene.

[0071] In this application, when determining candidate scenes, in order to reduce the degree of perception of scene switching on users, scenes similar to the target scene mode can be screened out as screening scenes. If the number of screening scenes is one, the screening scene is directly used as a candidate scene; if the number of screening scenes is greater than one, the screening scene with the highest activity level is used as a candidate scene based on the user's activity level in each screening scene.

[0072] For example, the screening scenes of scene A include scene B, scene C, and scene D. If the current user has the most visits to scene B among scene B, scene C, and scene D, scene B can be used as a candidate scene for scene A. In this way, the application can avoid using scenes with low activity as candidate scenes as much as possible, improve the video recommendation effect, and enhance the user experience.

[0073] In summary, the present application needs to pre-set candidate scenarios corresponding to each scenario, so that when the user accesses an inactive scenario, videos can be recommended through the active candidate scenarios to improve the recommendation effect. Moreover, when determining the candidate scenarios in the present application, in order to reduce the user's perception, it is necessary to screen out scenarios with similar modes to each scenario as candidate scenarios, reduce the perception brought by switching the scenario model to the user, and improve the user experience; if the number of scenarios with similar modes screened out is large, the candidate scenarios can be determined according to the activity level of the user in each screened scenario, so as to recommend more accurate videos for the user through the candidate scenarios with higher activity levels.

[0074] See Figure 3 , which is a schematic flowchart of another video recommendation method provided by an embodiment of the present application. The method specifically includes the following steps:

[0075] S301. Determine the user types of each user in different scenarios; wherein, the user types include: active users and inactive users;

[0076] S302. Receive an access instruction triggered by the current user; wherein, the access instruction is an instruction for the current user to access the target scenario;

[0077] S303. If the current user is an inactive user in the target scenario, determine the candidate scenario corresponding to the target scenario; the candidate scenario is a scenario with a mode similar to the target scenario;

[0078] S304. If the current user is an active user in the candidate scenario, recommend videos for the current user through the recommendation model of the candidate scenario.

[0079] In this embodiment, the user types of each user in different scenarios can be determined in advance according to the activity level information of the user in different scenarios; the user types include: active users and inactive users. The activity level information in the present application is to identify the activity level of the user in different scenarios through the user's historical behavior data, which can be the number of accesses and access time of the user to a certain scenario, etc., and is not specifically limited herein.

[0080] In another embodiment of the present application, the process of determining the user types of each user in different scenarios specifically includes: determining the number of accesses of each user in different scenarios; the number of accesses is the number of times each user accesses different scenarios within a predetermined time period; determining the user types of the users in different scenarios according to the number of accesses of each user in different scenarios and a predetermined threshold.

[0081] It should be noted that in this application, the access times of the user in different scenarios can represent the activity level of the user in different scenarios. In this embodiment, the user type can be divided into active users and inactive users according to the access times of the user accessing each scenario and a predetermined threshold. For example, if the predetermined threshold is set to 70, and the number of times the user accesses a certain scenario within a period of time is greater than 50, it is determined that the user is an active user in this scenario; otherwise, it is determined that the user is an inactive user in this scenario. Or, if the predetermined threshold is set to 10, and the number of times the user accesses a certain scenario within a period of time is less than 10, it is determined that the user is an inactive user in this scenario; otherwise, it is determined that the user is an active user in this scenario.

[0082] In summary, it can be seen that this application can pre-determine the user type of the user in different scenarios. This user type includes active users and inactive users. After determining the user type, when the user accesses the scenario, it can quickly determine whether it is necessary to switch the scenario model, so as to quickly and accurately recommend videos for the user. When determining the user type in this application, it can be determined according to the access times of the user accessing each scenario and a predetermined threshold. If the number of times the user accesses a certain scenario is large, it is determined as an active user. If the number of times the user accesses a certain scenario is small, it is determined that the user is an inactive user, thereby accurately determining the user type to more accurately recommend videos for the user.

[0083] In another embodiment of this application, after determining the user type of each user in different scenarios, it further includes:

[0084] According to the user type of each user in different scenarios, determine the target users corresponding to each scenario; where the target users are: the user type in each scenario is an inactive user, and the user type in the corresponding candidate scenario is an active user;

[0085] Correspondingly, if the current user is an active user in the candidate scenario, the process of recommending videos for the current user through the recommendation model of the candidate scenario includes: if the current user is the target user of the target scenario, it is determined that the current user is an active user in the candidate scenario, and through the recommendation model of the candidate scenario, recommend videos for the current user.

[0086] In this embodiment, after pre-determining the user type of the user in different scenarios, in order to quickly respond to the access instruction triggered by the current user, it is possible to pre-determine the users who need to switch the recommendation model in each scenario. The users who need to switch the recommendation model are: those who are inactive users in the current scenario but active users in the candidate scenario. Therefore, in this application, after determining the user type of each user in different scenarios, it is also necessary to determine the target users corresponding to each scenario. The target users are: the user type in each scenario is an inactive user, and the user type in the corresponding candidate scenario is an active user.

[0087] After setting the target user in this application, the access instruction triggered by the current user can be quickly responded to. For example: after this application receives the access instruction for the target scenario triggered by the current user, it can directly check whether the current user is the target user of the target scenario. If the current user is not the target user of the current scenario, it is determined that the current user is an active user in the target scenario, and in this case, there is no need to switch the recommendation model; if the current user is the target user of the current scenario, it is determined that the current user is an inactive user in the target scenario, and the current user is an active user in the candidate scenario. In this case, it is necessary to switch the recommendation model from the recommendation model of the target scenario to the recommendation model of the candidate scenario, and recommend videos for the current user through the recommendation model of the candidate scenario.

[0088] In summary, after determining the user types of the user in different scenarios in this application, the target users of each scenario can be directly determined. The target user is an inactive user in the current scenario and an active user in the candidate scenario. Therefore, when this application recommends videos, it can first detect whether the current user is the target user of the target scenario. If the current user is the target user of the target scenario, it can be directly determined to recommend videos through the candidate scenario, thereby improving the video recommendation speed.

[0089] See Figure 4 , which is another schematic flowchart of the video recommendation method provided by the embodiment of this application. The method specifically includes the following steps:

[0090] S401. Receive the access instruction triggered by the current user; where the access instruction is an instruction for the current user to access the target scenario;

[0091] S402. If the current user is an inactive user in the target scenario, determine the candidate scenario corresponding to the target scenario; the candidate scenario is a scenario with a similar pattern to the target scenario;

[0092] S403. If the current user is an active user in the target scenario, recommend videos for the current user through the recommendation model of the target scenario;

[0093] S404. If the current user is an active user in the candidate scenario, recommend videos for the current user through the recommendation model of the candidate scenario;

[0094] S405. If the current user is an inactive user in the candidate scenario, recommend videos for the current user through the recommendation model of the target scenario.

[0095] In this embodiment, after receiving an access instruction for the current user to access the target scenario, in order to improve the accuracy of video recommendation, the user type of the current user in the target scenario can be determined first. If the user type is an active user, it indicates that the user is relatively active in the target scenario. Therefore, the recommendation model of the target scenario can collect more user behavior data of the user, so the recommendation model of the target scenario can accurately recommend videos for the user, and there is no need to switch the recommendation model at this time. If the user type is an inactive user, it indicates that the user is not active in the target scenario. At this time, the recommendation model of the target scenario cannot collect more user behavior data, and the recommendation model of the target scenario may not be able to accurately recommend videos for the user.

[0096] In this case, the candidate scenarios of the target scenario can be determined, and the user type of the current user in the candidate scenarios can be determined. If the user type is an inactive user, it indicates that the user is not active in the candidate scenario. Since the user is an inactive user in both the target scenario and the candidate scenario, switching the recommendation model at this time cannot optimize the user experience, and there is no need to switch the recommendation model at this time. If the user type is an active user, it indicates that the user is relatively active in the candidate scenario. In this case, although the candidate scenario is different from the target scenario, since the patterns of the candidate scenario and the target scenario are similar, the user does not have a deep perception experience when switching scenarios, so the recommendation result of the user can be improved and the user experience can be optimized.

[0097] See Figure 5 , which is a schematic diagram of a recommendation result generation process provided by an embodiment of the present application. In this embodiment, two scenarios are set as: Scenario A and Scenario B; as shown in the figure, when the user accesses Scenario A, it is determined whether the user is an active user in Scenario A; if so, directly access the sorting model of Scenario A to obtain the recommendation result of Model A; if the user is an inactive user in Scenario A, it is determined whether the user is an active user in the candidate Scenario B; if the user is an inactive user in Scenario B, access the sorting model of Scenario A to obtain the recommendation result of Model A; if the user is an active user in Scenario B, access the sorting model of Scenario B to obtain the recommendation result of Model B. See Figure 6 , which is a schematic diagram of the overall framework of a recommendation result generation process provided by an embodiment of the present application; as shown in the figure, user behavior data will be generated when the user accesses the scenario. After processing data such as user behavior data, user portraits, and video features, training samples are generated, and the recommendation model of the scenario is trained through these training samples to obtain the trained recommendation model of the scenario; when the user accesses the scenario again, the recommendation model of the scenario can be used to recommend videos for the user. It should be noted that in order to improve the accuracy of video recommendation, the generation process of the training samples and the training process of the model are continuously ongoing.

[0098] In summary, in this application, if the user is an active user in the target scenario, the model of the target scenario can be directly used to recommend videos for the user, so as to accurately recommend videos for the user; if the user is an inactive user in the candidate scenario, switching the recommendation model may not improve the user experience. Therefore, in this case, in order to save the system resource consumption caused by switching the recommendation model, the model of the target scenario can be directly used to recommend videos for the user.

[0099] See Figure 7 , Figure 7 which is a schematic structural diagram of a video recommendation device provided by an embodiment of this application. The device specifically includes:

[0100] A receiving module 11, configured to receive an access instruction triggered by the current user; wherein, the access instruction is an instruction for the current user to access the target scenario;

[0101] A first determination module 12, configured to determine a candidate scenario corresponding to the target scenario when the current user is an inactive user in the target scenario; wherein, the candidate scenario is a scenario with a similar pattern to the target scenario;

[0102] A first recommendation module 13, configured to recommend videos for the current user through the recommendation model of the candidate scenario when the current user is an active user in the candidate scenario.

[0103] As an optional embodiment, the video recommendation device further includes:

[0104] A second determination module, configured to determine candidate scenarios corresponding to each scenario.

[0105] As an optional embodiment, the second determination module is specifically configured to: determine a screening scenario with a similar pattern to each scenario; if the number of screening scenarios is one, use the screening scenario as the candidate scenario; if the number of screening scenarios is greater than one, determine the activity level of the user in each screening scenario, and use the screening scenario with the highest activity level as the candidate scenario.

[0106] As an optional embodiment, the video recommendation device further includes:

[0107] A second recommendation module, configured to recommend videos for the current user through the recommendation model of the target scenario when the current user is an active user in the target scenario;

[0108] A third recommendation module, configured to recommend videos for the current user through the recommendation model of the target scenario when the current user is an inactive user in the candidate scenario.

[0109] As an optional embodiment, the video recommendation device further includes:

[0110] A third determination module, configured to determine the user types of each user in different scenarios; wherein, the user types include: active users and inactive users.

[0111] As an optional embodiment, the third determination module includes:

[0112] A first determination subunit, configured to determine the number of accesses of each user in different scenarios; wherein, the number of accesses is: the number of times each user accesses different scenarios within a predetermined time period;

[0113] A second determination subunit, configured to determine the user type of the user in different scenarios according to the number of accesses of each user in different scenarios and a predetermined threshold.

[0114] As an optional embodiment, the third determination module further includes:

[0115] A third determination subunit, configured to determine the target users corresponding to each scenario according to the user types of each user in different scenarios; wherein, the target users are: the user types in each scenario are inactive users, and the user types in the corresponding candidate scenarios are active users;

[0116] Correspondingly, the first recommendation module is specifically configured to: if the current user is the target user of the target scenario, determine that the current user is an active user in the candidate scenario, and recommend videos for the current user through the recommendation model of the candidate scenario.

[0117] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0118] See Figure 8 , Figure 8 , which is a schematic structural diagram of an electronic device provided by an embodiment of the present application. The electronic device specifically includes:

[0119] A processor 21, a memory 22, and a computer program stored on the memory 22 and executable on the processor 21. The processor 21 executes the steps of the video recommendation method described in any of the above method embodiments through the computer program.

[0120] Among them, the processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 may further include an AI (Artificial Intelligence) processor, and the AI processor is used to process computational operations related to machine learning.

[0121] The memory 22 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 22 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In this embodiment, the memory 22 is at least used to store the following computer program 221. After the computer program is loaded and executed by the processor 21, it can implement the relevant steps in the video recommendation method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 22 may further include an operating system 222 and data 223, etc., and the storage method may be temporary storage or permanent storage. Among them, the operating system 222 may include Windows, Unix, Linux, etc.

[0122] In some embodiments, the electronic device may further include a display screen 23, an input / output interface 24, a communication interface 25, a sensor 26, a power supply 27, and a communication bus 28.

[0123] Of course, Figure 8 The structure of the electronic device shown does not constitute a limitation on the electronic device in the embodiments of the present application. In practical applications, the electronic device may include more or fewer components than Figure 8 shown, or combine certain components.

[0124] In another exemplary embodiment, a computer storage medium is further provided. When the program instructions are executed by a processor, the steps of the video recommendation method described in any of the above method embodiments are implemented. Among them, the storage medium may include: various media that can store program codes such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0125] Optionally, specific examples in this embodiment may refer to the examples described in the above embodiments, and will not be elaborated herein.

[0126] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" as used herein may also include the plural forms. The terms "comprising", "including", "containing", and "having" are inclusive and thus specify the presence of the stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not to be construed as necessarily requiring them to be executed in the particular order described or illustrated, unless the execution order is explicitly stated. It should also be understood that alternative or additional steps may be used.

[0127] The above description is only the specific implementation manners of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather will be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A video recommendation method, characterized in that: include: Receive an access instruction triggered by the current user; wherein the access instruction is an instruction for the current user to access a target scene; If the current user is an inactive user in the target scene, determining a candidate scene corresponding to the target scene; wherein the candidate scene is a scene with a similar pattern to the target scene; If the current user is an active user in the candidate scene, a video is recommended for the current user through a recommendation model of the candidate scene.

2. The video recommendation method according to claim 1, characterized in that: Before receiving the access instruction triggered by the current user, the method further includes: A candidate scene corresponding to each scene is determined.

3. The video recommendation method according to claim 2, characterized in that: The determining of candidate scenes similar to each scene mode includes: Identify screening scenarios that are similar to the patterns of each scenario; If the number of screening scenarios is one, the screening scenario is taken as a candidate scenario; If the number of the screening scenarios is greater than one, the activity level of the user in each screening scenario is determined, and the screening scenario with the highest activity level is used as a candidate scenario.

4. The video recommendation method according to claim 1, characterized in that: After receiving the access instruction triggered by the current user, the method includes: If the current user is an active user in the target scene, recommending videos to the current user through a recommendation model of the target scene; If the current user is an inactive user in the candidate scene, a video is recommended for the current user through the recommendation model of the target scene.

5. The video recommendation method according to any one of claims 1 to 4, characterized in that: Before receiving the access instruction triggered by the current user, the method further includes: Determine the user type of each user in different scenarios; wherein the user type includes: active users and inactive users.

6. The video recommendation method according to claim 5, characterized in that: Determining the user type of each user in different scenarios includes: Determine the number of visits of each user in different scenarios; wherein the number of visits is: the number of visits of each user to different scenarios within a predetermined time period; The user type of the user in different scenarios is determined according to the number of visits of each user in different scenarios and a predetermined threshold.

7. The video recommendation method according to claim 5, characterized in that: After determining the user type of each user in different scenarios, the method further includes: Determine the target user corresponding to each scenario according to the user type of each user in different scenarios; wherein the target user is: an inactive user in each scenario and an active user in the corresponding candidate scenario; Correspondingly, if the current user is an active user in the candidate scene, recommending a video to the current user through the recommendation model of the candidate scene includes: If the current user is the target user of the target scene, it is determined that the current user is an active user in the candidate scene, and a video is recommended to the current user through a recommendation model of the candidate scene.

8. A video recommendation device, characterized in that: include: A receiving module, used to receive an access instruction triggered by the current user; wherein the access instruction is an instruction for the current user to access the target scene; A first determination module, configured to determine a candidate scene corresponding to the target scene when the current user is an inactive user in the target scene; wherein the candidate scene is a scene with a similar pattern to the target scene; The first recommendation module is used to recommend videos to the current user by using the recommendation model of the candidate scene when the current user is an active user in the candidate scene.

9. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the steps of the video recommendation method described in any one of claims 1 to 7 of the present application through the computer program.

10. A computer storage medium, characterized in that: The computer storage medium stores computer executable instructions, and the computer executable instructions are used to execute the steps of the video recommendation method described in any one of claims 1 to 7 of the present application.

Images