Information recommendation method, system and device

By judging user behavior data and user groups and supplementing candidate recommendation information, the problem of information recommendation for new users or users with fewer behaviors is solved, and the user experience and recommendation quality is improved.

CN113139122BActive Publication Date: 2024-11-22ALIBABA GROUP HOLDING LTD
View PDF 12 Cites 0 Cited by

Patent Information

Application Number
CN202010067079.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-01-20
Publication Date
2024-11-22
Estimated Expiration
2040-01-20

AI Technical Summary

Technical Problem

The existing information recommendation system is difficult to quickly find suitable recommended content in the case of new users or users with fewer behaviors, resulting in poor user experience, and the maintenance cost based on hot spots or manual compilation strategies is high and the quality is difficult to guarantee.

Method used

By determining whether the user behavior data meets the conditions for obtaining candidate recommendation information, if not, the candidate recommendation information is determined based on the user group to which the user belongs, and supplemented it into the candidate recommendation list to ensure the sustainability and quality of information recommendation.

Benefits of technology

It improves the quality of information recommendation for new users or users with fewer behaviors, improves the user experience, avoids interruption of information recommendation, and enhances the hit rate of user interest.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113139122B_ABST
    Figure CN113139122B_ABST
Patent Text Reader

Abstract

The embodiments of the present application provide an information recommendation method, system and device. The method includes the following steps: judging whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently met according to the user's behavior data; when judging that the first condition is not currently met, determining at least one first candidate recommendation information according to the user group to which the user belongs; and recommending information to the user according to the at least one first candidate recommendation information. The technical solution provided in the embodiments of the present application can obtain information based on the user group to which the user belongs when there is a lack of user behavior data, thereby solving the problem of insufficient information acquisition based on user behavior data; and supplementing information acquisition based on user groups can increase the probability of hitting user interests and improve user experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of information technology, and in particular to an information recommendation method, system and device. Background Art

[0002] With the development of Internet technology, the Internet can provide users with more and more network services, such as browsing videos, listening to music, reading, shopping, etc. On the Internet platform, users can search for the information they need through the search function. At the same time, in order to facilitate users to obtain information, the Internet platform can also actively recommend information to users.

[0003] Currently, a more popular information recommendation method is to recommend corresponding information to users based on their historical behavior data. Summary of the invention

[0004] In view of the above problems, the present application is proposed to provide an information recommendation method, system and device that solve the above problems or at least partially solve the above problems.

[0005] Therefore, in one embodiment of the present application, an information recommendation method is provided. The method comprises:

[0006] According to the user's behavior data, determining whether a first condition for obtaining candidate recommendation information based on the user's behavior data is currently met;

[0007] When it is determined that the first condition is not currently satisfied, determining at least one first candidate recommendation information according to the user group to which the user belongs;

[0008] Recommend information to the user based on the at least one first candidate recommendation information.

[0009] In one embodiment of the present application, an information recommendation system is provided. The system includes:

[0010] The client is used to send the user's behavior data to the server;

[0011] The server is used to determine, based on the behavior data, whether a first condition for obtaining candidate recommendation information based on the behavior data of the user is currently met; when it is determined that the first condition is not currently met, determine at least one first candidate recommendation information based on the user group to which the user belongs; and send the at least one candidate recommendation information to the client, so that the client recommends information to the user based on the at least one first candidate recommendation information.

[0012] In another embodiment of the present application, an information recommendation method is provided. The method comprises:

[0013] When there are not many candidate recommendation information in the candidate recommendation list corresponding to the user, determining at least one first candidate recommendation information according to the user group to which the user belongs;

[0014] The at least one first candidate recommendation information is added to the candidate recommendation list so as to continuously recommend information to the user.

[0015] In another embodiment of the present application, an information recommendation system is provided. The system includes:

[0016] The server is configured to obtain at least one first candidate recommendation information according to the user group to which the user belongs when the number of candidate recommendation information in the candidate recommendation list corresponding to the user is insufficient;

[0017] The client is used to receive the at least one first candidate recommendation information sent by the server; and add the at least one first candidate recommendation information to the candidate recommendation list, so as to continue to recommend information to the user.

[0018] In one embodiment of the present application, an electronic device is provided. The electronic device includes: a memory and a processor, wherein:

[0019] The memory is used to store programs;

[0020] The processor is coupled to the memory and is configured to execute the program stored in the memory to:

[0021] According to the user's behavior data, determining whether a first condition for obtaining candidate recommendation information based on the user's behavior data is currently met;

[0022] When it is determined that the first condition is not currently satisfied, determining at least one first candidate recommendation information according to the user group to which the user belongs;

[0023] Recommend information to the user based on the at least one first candidate recommendation information.

[0024] In another embodiment of the present application, an electronic device is provided. The electronic device includes: a memory and a processor, wherein:

[0025] The memory is used to store programs;

[0026] The processor is coupled to the memory and is configured to execute the program stored in the memory to:

[0027] When there are not many candidate recommendation information in the candidate recommendation list corresponding to the user, determining at least one first candidate recommendation information according to the user group to which the user belongs;

[0028] The at least one first candidate recommendation information is added to the candidate recommendation list so as to continuously recommend information to the user.

[0029] The technical solution provided in the embodiment of the present application determines whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently met based on the user's behavior data. When the first condition is not met, information is obtained based on the user group to which the user belongs, so as to recommend information to the user. In this way, when it is determined that there is a lack of user behavior data, information can be obtained based on the user group to which the user belongs, solving the problem of insufficient information obtained based on user behavior data; and supplementing information obtained based on user groups can increase the probability of hitting the user's interests and improve the user experience.

[0030] The technical solution provided in the embodiment of the present application can obtain information based on the user group to which the user belongs when candidate recommendation information is lacking, thereby avoiding interruptions in information recommendation; and supplementing information acquisition based on user groups can increase the probability of hitting the user's interests and enhance the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0032] Figure 1a A flowchart of a song recommendation method based on a smart speaker provided in one embodiment of the present application;

[0033] Figure 1b A flowchart of an information recommendation method provided by another embodiment of the present application;

[0034] Figure 2 A structural block diagram of an information recommendation system provided by another embodiment of the present application;

[0035] Figure 3 A flowchart of an information recommendation method provided by another embodiment of the present application;

[0036] Figure 4 A structural block diagram of an information recommendation device provided by another embodiment of the present application;

[0037] Figure 5 A structural block diagram of an information recommendation device provided by another embodiment of the present application;

[0038] Figure 6A structural block diagram of an electronic device provided in another embodiment of the present application;

[0039] Figure 7 A flowchart of an information recommendation method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0040] Taking music recommendation service as an example, many electronic devices, such as smartphones and smart speakers, can provide music recommendation services. Music recommendation service is one of the most important skills in smart speakers. The quality of user experience is directly related to the user's overall favorability towards smart speakers. The interaction method between smart speakers and APP (Application) on mobile phones is different, and the process of users' understanding or proficiency in using a specific skill of the speaker is relatively slow. And taking music listening as an example, each time a recommendation is made, only the content that is most likely to match the user's preferences can be recommended. In traditional recommendation schemes, there are two main links: information acquisition and information sorting. The first link is to find content that may be of interest based on the user's historical behavior data; the second link is to sort the content that may be of interest based on the user's long-term or short-term behavior and content characteristics, and recommend the content that the user likes.

[0041] In the process of implementing the technical solution provided by the embodiment of the present application, the inventors have found that the existing recommendation mechanism often fails to obtain sufficient information when facing new users, users who actively order songs less, and users who listen casually. In order to avoid the lack of recommended content, the system can use global hot content or manually sorted content as a backup. However, the inventors have found through actual observation that users often show great dissatisfaction with hot content; there are fewer manually sorted contents, the maintenance cost is high, and it may not be well suited to the tastes of different groups of people. However, in the music recommendation system, only the content that is most likely to match the user's preferences can be recommended each time, and there is no more opportunity for trial and error. It is necessary to ensure high quality as much as possible for each recommendation, which makes it difficult for this backup strategy to ensure the quality of each recommendation. It can be seen that the user experience is not very good when using hot data as a backup. Especially for new users, whether the initial recommendation is appropriate will directly affect the user's first impression of the product. The technical solution provided by the embodiment of the present application aims to solve the problem of being able to quickly find suitable music recommendation content and provide user experience when new users have no behavior or less user behavior.

[0042] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below according to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0043] In addition, some of the processes described in the specification, claims and the above-mentioned figures of the present application include multiple operations that appear in a specific order, and these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0044] Figure 1b A flow chart of an information recommendation method provided by an embodiment of the present application is shown. The executor of the method may be a client or a server. The client may be a hardware with an embedded program integrated on a terminal, or an application software installed in the terminal, or a tool software embedded in the terminal operating system, etc., which is not limited in the embodiment of the present application. The terminal may be any terminal device including a mobile phone, a tablet computer, a smart speaker, etc. The server may be a common server, a cloud or a virtual server, etc., which is not specifically limited in the embodiment of the present application. Figure 1b As shown, the method includes:

[0045] 101. Determine, based on user behavior data, whether a first condition for obtaining candidate recommendation information based on the user behavior data is currently satisfied.

[0046] 102. When it is determined that the first condition is not currently satisfied, determine at least one first candidate recommendation information according to the user group to which the user belongs.

[0047] 103. Recommend information to the user based on the at least one first candidate recommendation information.

[0048] In the above 101, in different application scenarios, the behavior types involved in the user's behavior data may also be different. For example: in the music recommendation scenario, the behavior types involved in the user's behavior data may include: listening, switching songs, on-demand, collecting, sharing, etc. One or more. For another example: in the video recommendation scenario, the behavior types involved in the user's behavior data may include: watching, on-demand, deleting, collecting, sharing, etc. One or more. For another example: in the news information recommendation scenario, the behavior types involved in the user's behavior data may include: browsing, searching, deleting, collecting, sharing, etc. One or more.

[0049] In one example, in a music recommendation scenario based on a smart speaker, the behavior types involved in the user's behavior data may include: listening, switching songs, on-demand, and collecting.

[0050] In one example, the number of behaviors in the user's behavior data can be used to determine whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently met. The number of behaviors is equal to the number of times the behavior occurs. Specifically, if the number of behaviors in the user's behavior data is less than or equal to the second preset threshold, it is determined that the first condition is not currently met. If the number of behaviors in the user's behavior data is greater than the second preset threshold, it is determined that the first condition is currently met. The size of the second preset threshold can be set according to actual needs, and this application does not make specific restrictions on this. In this embodiment, when the number of historical behaviors of the user is small, it will be determined that the first condition is not met.

[0051] In actual application, it is found that when obtaining information based on the user's behavior data, in order to improve the effectiveness of the acquisition, only certain specified behavior types will be used, and other behavior types will be ignored. Taking the music recommendation scenario as an example, the specified behavior types can be specifically: on-demand, collection, and sharing. Therefore, in another example, it can be determined whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently met based on the number of behaviors of the specified behavior type in the user's behavior data. Specifically, if the number of behaviors of the specified behavior type in the user's behavior data is less than or equal to the first preset threshold, it is determined that the first condition is not currently met. If the number of behaviors of the specified behavior type in the user's behavior data is greater than the first preset threshold, it is determined that the first condition is currently met. Among them, the behaviors of the specified behavior type can be specifically summarized as favorable behaviors, and favorable behaviors can accurately reflect the user's interests. Specifically, the above 101 "based on the user's behavior data, determine whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently met" can be implemented in the following steps:

[0052] 1011. Determine the number of favorable behaviors in the behavioral data.

[0053] 1012. When the number is less than a first preset threshold, it is determined that the first condition is not currently met.

[0054] The first preset threshold value can be set according to actual needs, and this application does not make any specific restrictions on this. The number of favorable behaviors in the above-mentioned behavior data is equal to the number of occurrences of favorable behaviors.

[0055] In actual application, user behaviors can be divided into favorable behaviors, unfavorable behaviors, and neutral behaviors. Among them, favorable behaviors can accurately reflect the user's interests and hobbies; unfavorable behaviors can accurately reflect the user's disgust; neutral behaviors can neither reflect the user's interests and hobbies nor the user's disgust. Taking the music recommendation scenario as an example, favorable behaviors may include: on-demand, collection, and sharing; unfavorable behaviors may include: switching songs; and neutral behaviors may include: listening to recommended music.

[0056] In the above 102, when it is determined that the first condition is not met, it means that it is currently impossible to obtain effective information based on the user's behavior data (i.e., information recall match), so it is necessary to obtain information based on the user group to which the user belongs, that is, to determine at least one first candidate recommendation information based on the user group to which the user belongs. Generally speaking, the interests and hobbies of users in a user group are relatively similar. Therefore, obtaining information based on the user group to which the user belongs can increase the probability of hitting the user's interests and improve the user experience.

[0057] In one example, the preference information of different user groups can be counted in advance, and then the preference information of the user group to which the user belongs can be directly obtained; at least one first candidate recommendation information is determined based on the preference information of the user group to which the user belongs. Part or all of the preference information of the user group to which the user belongs can be used as at least one first candidate recommendation information.

[0058] Specifically, multiple users can be grouped according to their user attributes to obtain multiple user groups; and the preference information of each user group can be counted according to the historical behavior data of each user group. Among them, each user group includes a first user group, and the first user group refers to any one of the user groups. Among them, the user attributes may include: one or more of gender, age, region, occupation, and place of origin. Specifically, users in the same region can be classified into the same user group, or users in the same place of origin can be classified into the same user group. Of course, multiple attributes of users can also be combined to group, so that the division of user groups is more refined and the probability of hitting user interests can be increased.

[0059] The above-mentioned “accumulating the preference information of the first user group according to the historical behavior data of the first user group” can be implemented by the following steps:

[0060] a. Count the first preference information of each user in the first user group.

[0061] b. De-duplication processing is performed on the first preference information of each user in the first user group to obtain a plurality of different second preference information.

[0062] c. Count the number of users in the first user group who prefer each of the second preference information.

[0063] d. Sort the multiple different second preference information according to the number of users.

[0064] c. According to the ranking result, determining the preference information of the first user group from a plurality of different second preference information.

[0065] For example: if the sorting result is a descending sorting result, a specified number of second preference information that are ranked first can be used as the preference information of the first user group. Since the first preference information of different users in the first user group is likely to be repeated, deduplication is required. In the above deduplication process, not only the deduplication of information with the same identity identifier is included, but also the deduplication of information with different identity identifiers but essentially the same content. Among them, the identity identifier can specifically be an ID (Identity document). For example: the song "***", due to comprehensive factors such as version, there may be multiple IDs in the song library that correspond to songs that are essentially "***".

[0066] In the process of counting the preference information of each user group based on the historical behavior data of users in each user group, filtering and adjusting operations on the historical behavior data of users may also be involved. Specifically: the historical behavior data of users in each user group may be filtered according to preset filtering conditions to filter out the valid historical behavior data of users in each user group; based on the valid historical behavior data of users in each user group, the preference information of each user group is counted. Among them, the filtering conditions can be set according to actual needs, and this application does not make specific restrictions on this. Taking the filtering condition as a time condition as an example, the filtering condition is the past month, that is, the historical behavior data within the past month is filtered out as the valid historical behavior data. Setting filtering conditions can effectively improve the effectiveness of statistics.

[0067] In the above 103, information is recommended to the user according to the at least one first candidate recommendation information. In some scenarios, one piece of information may be recommended to the user each time. For example, in music recommendation, only one piece of music is played each time.

[0068] The technical solution provided in the embodiment of the present application determines whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently met based on the user's behavior data. When the first condition is met, information is obtained based on the user group to which the user belongs, so as to recommend information to the user. In other words, when it is determined that there is a lack of user behavior data, information can be obtained based on the user group to which the user belongs, solving the problem of insufficient information obtained based on user behavior data; and supplementing information acquisition based on user groups can increase the probability of hitting the user's interests and improve the user experience.

[0069] In actual applications, when the number of favorable behaviors in the above behavior data is less than the first preset threshold, information acquisition will be performed based on the user group to which the user belongs. However, the small number of favorable behaviors in the above behavior data is also very important information, so information acquisition can be supplemented based on these small number of favorable behaviors. Specifically, the above method may also include:

[0070] 104. When the number is greater than 1 and less than the first preset threshold, determine at least one second candidate recommendation information according to the favorable behavior.

[0071] The number refers to the number of favorable behaviors in the above-mentioned behavior data, which is equal to the number of occurrences of favorable behaviors. In this embodiment, the first preset threshold is greater than 1. The favorable behaviors can be analyzed to obtain the user's interests; according to the user's interests, at least one second candidate recommendation information is matched. The specific analysis process and matching process can be referred to the prior art, which will not be repeated here.

[0072] Correspondingly, in the above 103, "recommending information to the user based on the at least one first candidate recommendation information" is specifically: recommending information to the user based on the at least one first candidate recommendation information and the at least one second candidate recommendation information.

[0073] Specifically, at least one first candidate recommendation information and at least one second candidate recommendation information may be added to a candidate recommendation list corresponding to the user, and information is recommended to the user based on the candidate recommendation list.

[0074] Considering that in actual applications, there are some such users: such users are not new users, and information has been obtained for them based on their favorable behaviors. For such users, when they have no new favorable behaviors or the newly generated favorable behaviors are few, and it is difficult to obtain sufficient information based on the newly generated favorable behaviors, the information can also be supplemented based on the user group of the user. In other words, "determining the number of favorable behaviors in the behavioral data" in the above 1011 is specifically: determining the number of newly generated favorable behaviors in the behavioral data. Among them, the newly generated favorable behaviors can be understood as favorable behaviors that have not yet been used to determine the candidate recommendation information.

[0075] Optionally, in the above 102, "when it is determined that the first condition is not currently met, at least one first candidate recommendation information is determined based on the user group to which the user belongs" is specifically: when it is determined that the first condition is not currently met and the number of existing candidate recommendation information corresponding to the user is insufficient, at least one first candidate recommendation information is determined based on the user group to which the user belongs.

[0076] In this way, for users who have been consuming the existing candidate recommendation information determined based on their past favorable behaviors and have not added new favorable behaviors, information can also be obtained and supplemented based on the user group to which they belong, so as to continuously recommend information to users and avoid the situation where there is no recommended content. In some application scenarios, such as music recommendation and short video recommendation, it is required to continuously recommend information to users, that is, to continuously play information to users.

[0077] In one example, the step of “recommending information to the user according to the at least one first candidate recommendation information” in 103 above can be implemented by:

[0078] 1031. Add the at least one first candidate recommendation information to the candidate recommendation list corresponding to the user.

[0079] 1032. Recommend information to the user according to the candidate recommendation list.

[0080] In an achievable solution, in the above 1031, the number of users in the user group who prefer each of the first candidate recommendation information can be obtained; based on the number of users, the ranking of at least one first candidate recommendation information in the candidate recommendation list is determined; and according to the ranking, at least one first candidate recommendation information is added to the candidate recommendation list. The larger the number of users of the first candidate recommendation information, the higher its ranking, and the higher it is recommended to the user. In the embodiment of the present application, recommending information to a user specifically refers to displaying the recommended information to the user.

[0081] In an achievable solution, in the above 1032, information may be recommended to the user according to the order of the information in the candidate recommendation list, and one piece of recommended information may be recommended to the user each time.

[0082] In order to be able to hit the user's interests as quickly as possible, the ranking of the information in the candidate recommendation list can also be adjusted based on the behavioral data generated by the user in the process of recommending information to the user. Specifically, the ranking of the information in the candidate recommendation list can be adjusted based on the favorable and unfavorable behaviors generated by the user in the process of recommending information to the user. For example: based on the user's unfavorable behavior, the content attributes that the user dislikes can be determined, and the ranking of the information in the candidate recommendation list related to the content attributes that the user dislikes can be lowered; based on the user's favorable behavior, the content attributes that the user prefers can be determined, and the ranking of the information in the candidate recommendation list related to the content attributes that the user prefers can be increased.

[0083] The ranking of information in the candidate recommendation list can be adjusted in real time. In existing recommendation systems, the ranking of information is usually adjusted based on the long-term behavior of users, but this ranking method basically loses the ranking effect for new users. At this time, a system based on real-time behavior data or short-term behavior data generated in the process of recommending information to the user can be designed to adjust the information ranking. In this way, user feedback can be perceived more sensitively, avoiding problems such as the overall recommendation effect gap caused by adjustment delays.

[0084] In order to further improve the probability of hitting the user's interests, the user's behavior in the process of recommending information to the user can also be combined to supplement the user's information acquisition. Specifically, the above method can also include:

[0085] 105. According to the user's disgusting behavior towards the recommended information during the process of recommending information to the user, obtain at least one third candidate recommended information that the user may not dislike.

[0086] 106. Supplement the at least one third candidate recommendation information to the candidate recommendation list corresponding to the user.

[0087] In the above 105, taking music recommendation as an example, the offensive behavior may specifically include song switching or offensive marking.

[0088] In one feasible solution, the above step 105 of “obtaining at least one third candidate recommendation information that the user may not dislike based on the user’s dislike behavior towards the recommended information during the process of recommending information to the user” can be implemented by the following steps:

[0089] 1051. According to the offensive behavior of the user generated in the process of recommending information to the user, determine the information that the user dislikes from the currently recommended information.

[0090] 1052. Determine attributes of the information that the user dislikes based on the information that the user dislikes.

[0091] 1053. Obtain the at least one third candidate recommendation information according to other information attributes among the multiple information attributes except the information attribute that the user dislikes.

[0092] In the above 1051, in one example, the information targeted by the offensive behavior can be determined based on the offensive behavior; and the information targeted by the offensive behavior is used as the information that the user is disgusted with.

[0093] Sometimes, the information that a user's offensive behavior targets is not necessarily information that the user dislikes. For example, in a music recommendation scenario, if a user switches songs when a song is about to end, this does not mean that the user dislikes the song. Therefore, in another example, the playback duration of the information that the offensive behavior targets can be determined based on the time point when the offensive behavior occurs and the playback start time point of the information that the offensive behavior targets; when the playback duration of the information that the offensive behavior targets is less than or equal to a preset duration threshold, the information that the offensive behavior targets is determined to be information that the user dislikes.

[0094] In the above 1052, the information that the user dislikes is analyzed to determine the attributes of the information that the user dislikes.

[0095] The information attributes may include but are not limited to: one or more of the creator, style, era, and language. Taking music recommendation as an example, the creator is specifically a singer; the style is specifically a music style or genre.

[0096] For example: among the offensive information, the information about creator A is the most, the information about style B is the most, the information about era C is the most, and the information about language D is the most. Then it can be determined that the information attributes that the user dislikes are creator A, style B, era C, and language D.

[0097] In the above 1053, the above multiple information attributes may be all information attributes set in advance. The at least one third candidate recommendation information is obtained according to other information attributes among the multiple information attributes except the information attribute that the user dislikes.

[0098] According to other information attributes, at least one third candidate recommendation information is obtained. At least one information corresponding to each content attribute in other information attributes can be obtained; according to at least one information corresponding to each content attribute in other information attributes, at least one third candidate recommendation information is determined. In one example, at least one third candidate recommendation information includes a specified number of information with a high popularity ranking among at least one information corresponding to each content attribute in other information attributes. The popularity of the information can be specifically determined based on the user level and / or the total historical playback time of the information. The larger the user level or the longer the total historical playback time, the higher the popularity and the higher the popularity ranking. Note: In the music recommendation and video recommendation scenarios, it can be determined based on the total historical playback time.

[0099] When acquiring information based on user groups, the information acquired for different users in the same user group is the same. When acquiring information based on the above-mentioned offensive behavior, the information acquired for different users in the same user group is different, and is closer to the interests of users than the information acquired based on user groups. Therefore, the step of acquiring information based on the above-mentioned offensive behavior is also called "differential acquisition." The combination of information acquisition based on user groups and information acquisition based on the generated offensive behavior realizes the recommendation of information that meets the interests of users from coarse to fine, thereby improving the user experience. In addition, combined with the behavioral data generated during the recommendation process, the order of information in the candidate recommendation list is adjusted, which can hit the interests of users more quickly, thereby accelerating the guidance to the mechanism of information acquisition and ordering based on behavioral data.

[0100] In practical applications, at least one piece of information matching different pre-set content attributes can be found in advance, and a corresponding relationship between each content attribute and the at least one piece of information matching it can be established, so that it can be directly obtained later.

[0101] It should be noted that for each content attribute, there may be duplicate information in the at least one information that matches it, so it is necessary to perform a deduplication operation on the at least one information that matches it, and obtain the at least one information that matches it after deduplication. Then, a corresponding relationship between each content attribute and the at least one information that matches it after deduplication is established. In the above deduplication process, not only does it include deduplication of information with the same identity identifier, but it also includes: deduplication of information with different identity identifiers but essentially the same content.

[0102] In order to capture the user's more fine-grained interests based on the aforementioned information acquisition based on user groups, the above 1053 "According to other information attributes in multiple information attributes except the information attribute that the user dislikes, obtain the at least one third candidate recommendation information" can be specifically: according to the other information attributes and the user attributes of the user, obtain at least one third candidate recommendation information. At least one information corresponding to each pairwise combination of each information attribute in the other information attributes and each user attribute of the user can be obtained; the at least one third candidate recommendation information includes a specified number of information with a top popularity ranking in the at least one information corresponding to each pairwise combination.

[0103] It is also possible to find at least one first information matching different combinations of content attributes and user attributes in advance; find first preference information of a user group with the user attributes in the combination; determine at least one information that is the same as the at least one first information and the first preference information; and establish a corresponding relationship between the combination and the at least one information. In this way, at least one information corresponding to each combination can be directly obtained later.

[0104] In the above 106, the at least one third candidate recommendation information may be directly added to the candidate recommendation queue.

[0105] In an achievable solution, the step of “adding the at least one third candidate recommendation information to the candidate recommendation list corresponding to the user” in 106 can be implemented by the following steps:

[0106] 1061. Determine the user's tolerance level for currently recommended information based on the offensive behavior.

[0107] 1062. When the tolerance level is less than or equal to a first preset level, add the at least one third candidate recommendation information to the candidate recommendation list.

[0108] In the above 1061, in one example, the frequency of occurrence of the offensive behavior can be counted, and the tolerance level can be determined based on the frequency of occurrence. The greater the frequency of occurrence, the lower the tolerance level. The frequency of occurrence refers to the number of occurrences per unit time. In another example, the information that the user dislikes can be determined based on the offensive behavior; the tolerance level can be determined based on the ratio of the number of information that the user dislikes to the total number of currently recommended information. The larger the ratio, the lower the tolerance level.

[0109] In the above embodiment, the user's tolerance is estimated through rule strategies. Of course, the user's tolerance can also be estimated through a trained algorithm model. Therefore, in another example, the "determine the user's tolerance for the currently recommended information based on the offensive behavior" in the above 1061 can be specifically: determine the input information of the trained computing model based on the offensive behavior; input the input information into the computing model to obtain the tolerance output by the computing model. According to the offensive behavior, the number of information targeted by the offensive behavior, the number of consecutive offensive behaviors, the number of consecutive non-offensive behaviors, and the average playback time of the currently recommended information can be determined as input information. The input information is input into the computing model, and the computing model predicts the user's tolerance for the currently recommended information based on the input information. In this embodiment, the computing model judges the user's tolerance based on multiple dimensional information, which can improve the accuracy of the judgment. And the computing model can effectively improve its prediction accuracy through continuous learning or training. The computing model can be trained in advance based on a large number of training samples, and a training sample can include a test user's disgusting behavior towards multiple test information and his or her actual tolerance for multiple test information feedback. The internal implementation of the computing model can be set according to actual needs, and this application does not make specific limitations on this.

[0110] In the above 1062, the size of the first preset degree can be set according to actual needs, and the embodiment of the present application does not specifically limit this. If the tolerance level is less than or equal to the first preset level, it means that the user obviously does not like the recommendation based on at least one first candidate recommendation information at this time, and the recommendation based on this range can be stopped.

[0111] Note: In the candidate recommendation list, once each candidate recommendation information is recommended to the user, the candidate recommendation information will be automatically deleted from the candidate recommendation list or marked as recommended.

[0112] Optionally, the above method may further include:

[0113] 107. Determine at least one fourth candidate recommendation information based on the favorable behavior of the user generated in the process of recommending information to the user.

[0114] 108. Add the at least one fourth candidate recommendation information to the candidate recommendation list corresponding to the user.

[0115] In the above 107, the user's interests and hobbies are determined based on the generated favorable behavior; and at least one fourth candidate recommendation information is matched and obtained based on the interests and hobbies.

[0116] In the above 108, the at least one fourth candidate recommendation information is added to the candidate recommendation list corresponding to the user.

[0117] In one example, the above method may further include:

[0118] 109. Obtain user attributes of the user.

[0119] 110. Determine the user group to which the user belongs based on the user attributes.

[0120] In the above 109, the user attribute can be prompted to enter by the user; or it can be obtained by obtaining the login account information of the user; or it can be obtained from the login account information. Among them, the login account can be an account registered by the user on the platform. Of course, many platforms now support users to log in using their accounts registered on other platforms, so the login account can also be an external account that the user has registered on other platforms. Usually, the login account information includes gender, age, region, occupation, place of origin, etc.

[0121] User attributes may include: one or more of: gender, age, region, occupation, and place of origin. In the music recommendation scenario, the first candidate recommendation information is the first candidate recommended music. In the above 103, "recommending information to the user based on the at least one first candidate recommended music" specifically means: playing recommended music to the user based on the at least one first candidate recommended music. Recommend and play one piece of recommended music to the user each time.

[0122] In the music recommendation scenario of the home version of the smart speaker, when the smart speaker plays music, usually multiple family members will listen to it. In the above embodiments, when obtaining candidate recommendation information based on the user group, it is completely based on the attribute information of a certain user, and other family members are not taken into account, and unreasonable recommendations may occur. For example, if there is a baby member in a family, it is best not to recommend rock-style or sad-style music. Therefore, in another embodiment, the at least one first candidate recommendation information obtained based on the user group to which the user belongs can be further screened in combination with family information to improve family satisfaction. Family information can be actively input by the user or obtained through voice or image recognition with the authorization of the user, and this application does not make specific restrictions on this. According to the family information, the first information style (for example, music style or genre) to be filtered out can be determined; the candidate recommendation information of the first information style in at least one first candidate recommendation information is filtered out to obtain at least one filtered candidate recommendation information. Subsequently, based on at least one filtered candidate recommendation information, it is recommended to the user.

[0123] In addition, in the music recommendation scenario, the music styles that users want to listen to in different time periods may also be different. For example, a user likes to listen to some dynamic music during the day, and wants to listen to some quiet music late at night. In order to meet this demand, the obtained candidate recommendation information can also be filtered in combination with the current time. Specifically, the style of information to be filtered out can be determined according to the current time; the obtained candidate recommendation information can be filtered. For example, if the current time is late at night, it is determined that the music styles to be filtered out are: rock style.

[0124] Figure 2 FIG. 1 shows a schematic diagram of the structure of an information recommendation system provided by an embodiment of the present application. Figure 2 As shown, the system includes: a client 10 and a server 20. Among them,

[0125] The client 10 is used to send the user's behavior data to the server 20;

[0126] The server 20 is used to determine, based on the behavior data, whether a first condition for obtaining candidate recommendation information based on the user's behavior data is currently met; when it is determined that the first condition is not currently met, determine at least one first candidate recommendation information based on the user group to which the user belongs; and send the at least one candidate recommendation information to the client, so that the client recommends information to the user based on the at least one first candidate recommendation information.

[0127] The client 10 may send the user's behavior data to the server 20 in real time or at a first preset time interval.

[0128] In addition, the client 10 is further configured to: send a recommendation request to the server in response to a trigger operation of the user. The server 20 is configured to, upon receiving the recommendation request sent by the client, determine, based on the behavior data, whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently satisfied.

[0129] Of course, the server 20 can automatically determine whether the first condition for obtaining candidate recommendation information based on the user's behavior data is currently satisfied according to the behavior data at every second preset time interval.

[0130] The client 101 may be a hardware with an embedded program integrated on the terminal, or an application software installed in the terminal, or a tool software embedded in the terminal operating system, etc., which is not limited in this embodiment. The terminal may be any device with a certain computing capability, such as a smart phone, a tablet computer, a personal computer, a smart speaker, etc. The basic structure of the terminal with the client 101 may include: at least one processing unit and at least one memory. The number of processing units and memories depends on the configuration and type of the terminal. The memory may include volatile, such as RAM, or non-volatile, such as read-only memory (ROM), flash memory, etc., or may include both types at the same time. The memory usually stores an operating system (OS), one or more application programs, and may also store program data, etc. In addition to the processing unit and memory, the terminal also includes some basic configurations, such as a network card chip, an IO bus, audio and video components, etc. Optionally, the terminal may also include some peripheral devices, such as a keyboard, a mouse, an input pen, etc. Other peripheral devices are well known in the art and will not be described in detail here.

[0131] The server 102 refers to a server that can provide information interaction services in a network environment. Users can send images to the server 102 through the network, and can also obtain target ellipse information from the server 102 through the network. In physical implementation, the server 102 can be any device that can provide computing services, respond to service requests, and perform processing, such as a conventional server, cloud server, cloud host, virtual center, etc. The composition of the server mainly includes a processor, a hard disk, a memory, a system bus, etc., which is similar to a general computer architecture.

[0132] In this embodiment, the client 101 is connected to the server 102. The client 101 and the server 102 may be connected via a wireless or wired network. If the client 101 is connected to the server 102 for communication, the network standard of the mobile network may be any one of 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G+ (LTE+), WiMax, 5G, etc.

[0133] The execution principles and interaction processes of the various components in the system embodiment, such as the client and the server, can refer to the descriptions of the above-mentioned method embodiments.

[0134] The following will take music recommendation in a smart speaker as an example to introduce the technical solution of the embodiment of the present application in detail:

[0135] like Figure 1aAs shown, after user A turns on the smart speaker 50, he enters the casual listening scene. The smart speaker 50 can send a prompt signal to the server 51 that the user has entered the casual listening scene. After receiving the prompt signal, the server 51 can perform the following steps in sequence:

[0136] Step 601: Obtain the behavior data of user A.

[0137] Step 602: Based on the behavior data of user A, determine whether the first condition for obtaining candidate recommendation information based on the behavior data of the user is currently met.

[0138] For example: the number of favorable behaviors in the behavior data of user A is obtained; when the number is less than or equal to a first preset threshold, it is determined that the first condition is currently met.

[0139] Step 603: When it is determined that the first condition is not currently satisfied, determine the user group to which user A belongs.

[0140] Step 604 : After determining that the user group to which user A belongs is user group B, the preference set K of user group B is returned to the smart speaker 50 .

[0141] After receiving the set K, the smart speaker 50 adds the information in the set K to the candidate playlist corresponding to the user A, and the smart speaker 50 plays the recommended information in sequence according to the candidate playlist. While listening to songs, user A can provide feedback through voice or manual triggering through the interface. For example: when user A does not like a recommended song, the user can send a voice command to the smart speaker 50 "please switch to the next song" to make the smart speaker 50 switch to the next song. Another example: when user A hears a song he likes, the user can send a voice command to the smart speaker 50 "please collect this song" to make the smart speaker 50 perform the collection operation.

[0142] In the process of playing recommended songs for users, the smart speaker 50 can send the user's feedback behavior to the server 51, so that the server performs subsequent operations according to the feedback behavior, such as Figure 7 As shown:

[0143] 701. Play songs for the user according to the candidate playlist.

[0144] 702. Determine whether the song-changing behavior is frequent.

[0145] When the number of song switching events is small, step 703 is executed; when the number of song switching events is large, step 704 is executed.

[0146] 703. Determine whether the number of candidate play songs in the candidate play list is small.

[0147] If the number is small, step 704 is executed, that is, difference acquisition is performed based on the song-changing behavior. The song-changing behavior is the above-mentioned offensive behavior.

[0148] The specific implementation of step 704 can refer to the corresponding content in the above embodiments, which will not be described in detail here. The server 51 can return the songs obtained by the difference to the smart speaker 50, so that the smart speaker 50 can add them to the candidate playlist.

[0149] In addition, the server can also sort the information in the candidate playlist based on the user's feedback behavior. The specific process can be found in the corresponding content of the above embodiments, which will not be repeated here.

[0150] After testing, it was found that the technical solution provided by the embodiment of the present application can increase the average online traffic time by 1.04% and reduce the song-changing rate by 2.55%. And statistics on the information acquisition distribution of new users activated on the same day in the casual listening scenario showed that the proportion of information acquisition based on user groups was about 60%.

[0151] Figure 3 A flow chart of an information recommendation method provided by another embodiment of the present application is shown. The execution subject of the method may be a client or a server. The client may be a hardware with an embedded program integrated on a terminal, an application software installed in the terminal, or a tool software embedded in a terminal operating system, etc., which is not limited in the embodiments of the present application. The terminal may be any terminal device including a mobile phone, a tablet computer, a smart speaker, etc. The server may be a common server, a cloud or a virtual server, etc., which is not specifically limited in the embodiments of the present application.

[0152] like Figure 3 As shown, the method includes:

[0153] 201. When there are not many candidate recommendation information in the candidate recommendation list corresponding to the user, determine at least one first candidate recommendation information according to the user group to which the user belongs.

[0154] 202. Add the at least one first candidate recommendation information to the candidate recommendation list, so as to continuously recommend information to the user.

[0155] In the above 201, the number of candidate recommendation information in the candidate recommendation list may be obtained; when the number of candidate recommendation information is less than or equal to the third preset threshold, it may be determined that the number of candidate recommendation information in the candidate recommendation list is small.

[0156] The candidate recommendation information already in the candidate recommendation list may be determined based on the user's behavior data.

[0157] The specific implementation of "determining at least one first candidate recommendation information according to the user group to which the user belongs" in the above 201 can be found in the corresponding content of the above embodiments, which will not be repeated here.

[0158] In the above 202, at least one first candidate recommendation information is added to the candidate recommendation list; and subsequently, information is recommended to the user according to the candidate recommendation list.

[0159] The technical solution provided in the embodiment of the present application can obtain information based on the user group to which the user belongs when candidate recommendation information is lacking, thereby avoiding interruptions in information recommendation; and supplementing information acquisition based on user groups can increase the probability of hitting the user's interests and enhance the user experience.

[0160] Optionally, the candidate recommendation list is a candidate play list. In the above 202, "adding the at least one first candidate recommendation information to the candidate recommendation list so as to continuously recommend information to the user" specifically means: adding the at least one first candidate recommendation information to the candidate play list so as to continuously play the recommended information to the user. When playing the recommended information to the user, only one recommended information is played each time.

[0161] It should be noted here that: for the contents of each step not fully described in detail in the method provided in the embodiment of the present application, please refer to the corresponding contents in the above embodiment, which will not be repeated here. In addition, in addition to the above steps, the method provided in the embodiment of the present application may also include other parts or all of the steps in the above embodiments, which can be specifically referred to the corresponding contents in the above embodiments, which will not be repeated here.

[0162] Figure 2 The structure diagram of the information recommendation system provided by an embodiment of the present application is shown. The system includes: a client 10 and a server 20.

[0163] The server 20 is configured to obtain at least one first candidate recommendation information according to the user group to which the user belongs when the number of candidate recommendation information in the candidate recommendation list corresponding to the user is insufficient;

[0164] The client 10 is used to receive the at least one first candidate recommendation information sent by the server 20; and add the at least one first candidate recommendation information to the candidate recommendation list, so as to continuously recommend information to the user.

[0165] In one example, the client 10 may send the number of candidate recommendation information in the candidate recommendation list corresponding to the user to the server 20 at every fourth preset time interval, so that the server 20 can determine whether the number of candidate recommendation information in the candidate recommendation list corresponding to the user is insufficient. In another example, the client 10 may determine whether the number of candidate recommendation information in the candidate recommendation list corresponding to the user is insufficient based on the number of candidate recommendation information in the candidate recommendation list corresponding to the user at every fifth preset time interval; when it is determined that the number of candidate recommendation information in the candidate recommendation list corresponding to the user is insufficient, a prompt message about the insufficient number is sent to the server 20, so as to trigger the server 20 to obtain at least one first candidate recommendation information according to the user group to which the user belongs.

[0166] The client 101 may be a hardware with an embedded program integrated on the terminal, or an application software installed in the terminal, or a tool software embedded in the terminal operating system, etc., which is not limited in this embodiment. The terminal may be any device with a certain computing capability, such as a smart phone, a tablet computer, a personal computer, a smart speaker, etc. The basic structure of the terminal with the client 101 may include: at least one processing unit and at least one memory. The number of processing units and memories depends on the configuration and type of the terminal. The memory may include volatile, such as RAM, or non-volatile, such as read-only memory (ROM), flash memory, etc., or may include both types at the same time. The memory usually stores an operating system (OS), one or more application programs, and may also store program data, etc. In addition to the processing unit and memory, the terminal also includes some basic configurations, such as a network card chip, an IO bus, audio and video components, etc. Optionally, the terminal may also include some peripheral devices, such as a keyboard, a mouse, an input pen, etc. Other peripheral devices are well known in the art and will not be described in detail here.

[0167] The server 102 refers to a server that can provide information interaction services in a network environment. Users can send images to the server 102 through the network, and can also obtain target ellipse information from the server 102 through the network. In physical implementation, the server 102 can be any device that can provide computing services, respond to service requests, and perform processing, such as a conventional server, cloud server, cloud host, virtual center, etc. The composition of the server mainly includes a processor, a hard disk, a memory, a system bus, etc., which is similar to a general computer architecture.

[0168] In this embodiment, the client 101 is connected to the server 102. The client 101 and the server 102 may be connected via a wireless or wired network. If the client 101 is connected to the server 102 for communication, the network standard of the mobile network may be any one of 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G+ (LTE+), WiMax, 5G, etc.

[0169] The execution principles and interaction processes of the various components in the system embodiment, such as the client and the server, can refer to the descriptions of the above-mentioned method embodiments.

[0170] Figure 4 FIG. 1 shows a structural block diagram of an information recommendation device provided by an embodiment of the present application. Figure 4 As shown, the device comprises:

[0171] A first judgment module 301 is used to judge whether a first condition for obtaining candidate recommendation information based on the user's behavior data is currently met according to the user's behavior data;

[0172] A first determination module 302, configured to determine at least one first candidate recommendation information according to the user group to which the user belongs when it is determined that the first condition is not currently satisfied;

[0173] The first recommendation module 303 is configured to recommend information to the user based on the at least one first candidate recommendation information.

[0174] Optionally, the first determining module 302 is further configured to:

[0175] When the number is greater than 1 and less than the first preset threshold, at least one second candidate recommendation information is determined according to the favorable behavior;

[0176] The first recommendation module 303 is specifically used for:

[0177] Recommend information to the user based on the at least one first candidate recommendation information and the at least one second candidate recommendation information.

[0178] Optionally, the above device further includes:

[0179] A second determination module is used to obtain at least one third candidate recommendation information that the user may not dislike based on the user's dislike behavior towards the recommended information during the process of recommending information to the user;

[0180] The second supplementing module is used to supplement the at least one third candidate recommendation information into the candidate recommendation list corresponding to the user.

[0181] Optionally, the first determining module is further used to:

[0182] Determine at least one fourth candidate recommendation information according to the favorable behavior of the user generated in the process of recommending information to the user;

[0183] The above device may further include:

[0184] The first adding module is used to add the at least one fourth candidate recommendation information to the candidate recommendation list corresponding to the user.

[0185] The above device may further include:

[0186] The first adjustment module is used to adjust the order of the information in the candidate recommendation list according to the behavior data generated by the user in the process of recommending information to the user.

[0187] The above device may further include:

[0188] A second acquisition module, used to acquire user attributes of the user;

[0189] The third determination module is used to determine the user group to which the user belongs according to the user attributes.

[0190] It should be noted here that: the information recommendation device provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles and beneficial effects of the above modules can be found in the corresponding contents of the above method embodiments, which will not be repeated here.

[0191] Figure 5 FIG. 1 shows a structural block diagram of an information recommendation device provided by an embodiment of the present application. Figure 5 As shown, the device comprises:

[0192] The fourth determining module 401 is used to determine at least one first candidate recommendation information according to the user group to which the user belongs when the number of candidate recommendation information in the candidate recommendation list corresponding to the user is small;

[0193] The first supplementing module 402 is used to supplement the at least one first candidate recommendation information into the candidate recommendation list, so as to continuously recommend information to the user.

[0194] Optionally, the above device further includes:

[0195] A fifth determining module, configured to obtain at least one third candidate recommended information according to the user's disgusting behavior towards the recommended information during the process of recommending information to the user;

[0196] The third supplementing module is used to supplement the at least one third candidate recommendation information into the candidate recommendation list corresponding to the user.

[0197] Optionally, the fourth determining module is further used to:

[0198] Determine at least one fourth candidate recommendation information according to the favorable behavior of the user generated in the process of recommending information to the user;

[0199] The above device further comprises:

[0200] The second adding module is used to add the at least one fourth candidate recommendation information to the candidate recommendation list corresponding to the user.

[0201] Optionally, the above device further includes:

[0202] The second adjustment module is used to adjust the order of the information in the candidate recommendation list according to the behavior data generated by the user in the process of recommending information to the user.

[0203] Optionally, the above device further includes:

[0204] A fourth acquisition module, used to acquire user attributes of the user;

[0205] The sixth determination module is used to determine the user group to which the user belongs according to the user attributes.

[0206] It should be noted here that: the information recommendation device provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles and beneficial effects of the above modules can be found in the corresponding contents of the above method embodiments, which will not be repeated here.

[0207] Figure 6 FIG. 1 is a schematic diagram showing the structure of an electronic device provided by an embodiment of the present application. Figure 6 As shown, the electronic device includes a memory 1101 and a processor 1102. The memory 1101 can be configured to store various other data to support operations on the electronic device. Examples of these data include instructions for any application or method for operating on the electronic device. The memory 1101 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk.

[0208] The memory 1101 is used to store programs;

[0209] The processor 1102 is coupled to the memory 1101 and is used to execute the program stored in the memory 1101 to implement the information recommendation method provided by the above-mentioned method embodiments.

[0210] Further, if Figure 6 As shown, the electronic device also includes: a communication component 1103, a display 1104, a power component 1105, an audio component 1106 and other components. Figure 6 Only some components are shown schematically, which does not mean that the electronic device only includes Figure 6 Components shown.

[0211] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a computer, the steps or functions of the information recommendation method provided by the above-mentioned method embodiments can be implemented.

[0212] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0213] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0214] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An information recommendation method, characterized in that: include: Determining, based on the user's behavior data, whether a first condition for obtaining candidate recommendation information based on the user's behavior data is currently satisfied, wherein the user is a user for whom candidate recommendation information has been determined and recommended based on the user's favorable behavior in the past, including: determining the number of newly generated favorable behaviors in the behavior data; when the number is less than a first preset threshold, determining that the first condition is not currently satisfied; wherein the newly generated favorable behaviors refer to favorable behaviors that have not yet been used to determine candidate recommendation information for the user; When it is determined that the first condition is not satisfied currently and the number of existing candidate recommendation information corresponding to the user that has not been recommended to the user is insufficient, at least one first candidate recommendation information is determined according to the user group to which the user belongs, and the existing candidate recommendation information corresponding to the user is determined based on the user's historical favorable behavior; Recommend information to the user one by one according to the at least one first candidate recommendation information.

2. The method according to claim 1, characterized in that: Also includes: When the number is greater than 1 and less than the first preset threshold, at least one second candidate recommendation information is determined according to the favorable behavior; Recommending information to the user according to the at least one first candidate recommendation information includes: Recommend information to the user based on the at least one first candidate recommendation information and the at least one second candidate recommendation information.

3. The method according to claim 1 or 2, characterized in that: Determining the number of favorable behaviors in the behavioral data includes: The number of newly generated favorable behaviors in the behavioral data is determined.

4. The method according to claim 3, characterized in that When it is determined that the first condition is not currently met, at least one first candidate recommendation information is determined according to the user group to which the user belongs, including: When it is determined that the first condition is not currently satisfied and the amount of existing candidate recommendation information corresponding to the user is insufficient, at least one first candidate recommendation information is determined according to the user group to which the user belongs.

5. The method according to any one of claims 1 to 2, characterized in that Determining at least one first candidate recommendation information according to the user group to which the user belongs, including: Obtaining preference information of the user group to which the user belongs; The at least one first candidate recommendation information is determined according to the preference information of the user group.

6. The method according to any one of claims 1 to 2, characterized in that Recommending information to the user according to the at least one first candidate recommendation information includes: Adding the at least one first candidate recommendation information to the candidate recommendation list corresponding to the user; Recommend information to the user based on the candidate recommendation list.

7. The method according to claim 6, characterized in that Also includes: According to the offensive behavior of the user generated in the process of recommending information to the user, obtaining at least one third candidate recommendation information that the user may not be offensive to; The at least one third candidate recommendation information is added to the candidate recommendation list corresponding to the user.

8. The method according to claim 7, characterized in that Supplementing the at least one third candidate recommendation information to the candidate recommendation list corresponding to the user includes: Determining the user's tolerance level for currently recommended information based on the offensive behavior; When the tolerance level is less than or equal to a first preset level, the at least one third candidate recommendation information is added to the candidate recommendation list.

9. The method according to claim 8, characterized in that Determining the user's tolerance level for currently recommended information based on the offensive behavior includes: Determining input information of a trained computing model based on the objectionable behavior; The input information is input into the calculation model to obtain the tolerance level output by the calculation model.

10. The method according to claim 7, characterized in that Acquiring at least one third candidate recommendation information that the user may not dislike according to the offensive behavior generated by the user in the process of recommending information to the user, including: Determining information that the user dislikes from currently recommended information based on the user's disliked behavior generated during the process of recommending information to the user; Determining, based on the information objectionable to the user, an attribute of the information objectionable to the user; The at least one third candidate recommendation information is obtained according to other information attributes among the multiple information attributes except the information attribute that the user dislikes.

11. The method according to claim 10, characterized in that Acquiring the at least one third candidate recommendation information according to other information attributes among the multiple information attributes except the information attribute that the user dislikes, includes: At least one third candidate recommendation information is obtained according to the other information attributes and the user attributes of the user.

12. The method according to claim 6, characterized in that Also includes: Determine at least one fourth candidate recommendation information according to the favorable behavior of the user generated in the process of recommending information to the user; The at least one fourth candidate recommendation information is added to the candidate recommendation list corresponding to the user.

13. The method according to claim 6, characterized in that Recommending information to the user according to the candidate recommendation list includes: Recommend information to the user according to the order of the information in the candidate recommendation list.

14. The method according to claim 13, characterized in that Also includes: The order of the information in the candidate recommendation list is adjusted according to the behavior data generated by the user in the process of recommending information to the user.

15. The method according to any one of claims 1 to 2, characterized in that Also includes: Obtaining user attributes of the user; The user group to which the user belongs is determined according to the user attributes.

16. The method according to claim 15, characterized in that Obtaining user attributes of the user, including: Obtaining the user's login account information; The user attributes of the user are obtained from the login account information.

17. The method according to any one of claims 1 to 2, characterized in that The first candidate recommendation information is the first candidate recommended music; Recommending information to the user according to the at least one first candidate recommendation information includes: Based on the at least one first candidate recommended music, recommended music is played for the user.

18. An information recommendation system, characterized in that: include: The client is used to send the user's behavior data to the server; The server is used to determine, based on the behavior data, whether a first condition for obtaining candidate recommendation information based on the behavior data of the user is currently met, wherein the user is a user for whom candidate recommendation information has been determined and recommended based on the user's favorable behavior in the past, including: determining the number of newly generated favorable behaviors in the behavior data; when the number is less than a first preset threshold, determining that the first condition is not currently met; wherein the newly generated favorable behaviors refer to favorable behaviors that have not yet been used to determine candidate recommendation information for the user; when it is determined that the first condition is not currently met and the number of existing candidate recommendation information corresponding to the user that has not yet been recommended to the user is insufficient, determining at least one first candidate recommendation information based on the user group to which the user belongs; and sending the at least one candidate recommendation information to the client, so that the client recommends information to the user one by one based on the at least one first candidate recommendation information, wherein the existing candidate recommendation information corresponding to the user is determined based on the user's favorable behavior in history.

19. An electronic device, characterized in that: include: A memory and a processor, wherein The memory is used to store programs; The processor is coupled to the memory and is configured to execute the program stored in the memory to: Determining, based on the user's behavior data, whether a first condition for obtaining candidate recommendation information based on the user's behavior data is currently satisfied, wherein the user is a user for whom candidate recommendation information has been determined and recommended based on the user's favorable behavior in the past, including: determining the number of newly generated favorable behaviors in the behavior data; when the number is less than a first preset threshold, determining that the first condition is not currently satisfied; wherein the newly generated favorable behaviors refer to favorable behaviors that have not yet been used to determine candidate recommendation information for the user; When it is determined that the first condition is not satisfied currently and the number of existing candidate recommendation information corresponding to the user that has not been recommended to the user is insufficient, at least one first candidate recommendation information is determined according to the user group to which the user belongs, and the existing candidate recommendation information corresponding to the user is determined based on the user's historical favorable behavior; Recommend information to the user one by one according to the at least one first candidate recommendation information.

Citation Information

Patent Citations

  • Method for recommending multi-media files and recommendation server

    CN104123325A

  • User behavior-based personalized audio recommendation method and system

    CN106326277A

  • Information recommendation method and device

    CN106326431A

  • Video recommendation method and system, and equipment

    CN107657004A

  • Music pushing method and speaker device

    CN108200142A