Recommendation method and device, electronic equipment and storage medium

Abstract

The invention relates to a recommendation method and device, electronic equipment and a storage medium, and the recommendation method comprises the steps: obtaining account information and equipment information of a user behavior to determine at least two equipment clusters corresponding to the account information, and determining the equipment cluster to which the equipment information belongs as a target equipment cluster; then, according to the historical behavior characteristics corresponding to the target device cluster, intention prediction is carried out in combination with the user behavior, recommendation information corresponding to the user behavior is generated, and therefore the recommendation information can be directly generated for the device cluster associated with the account information and the device information; therefore, the problem that in the prior art, only the user behavior corresponding to the account serves as the basis of personalized recommendation, and accurate personalized recommendation cannot be carried out due to disordered user behaviors can be avoided, recommendation information can be accurately and adaptively generated for each user, and recommendation accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of computers, and in particular to a recommendation method, device, electronic device and storage medium. Background Art

[0002] With the development of the Internet, operators such as various applications or websites usually make personalized recommendations for users in order to improve user stickiness and user experience. The existing recommendation method is to recommend corresponding content to users based on user personal portraits and behavioral habits. The user personal portraits and behavioral habits used are usually collected and analyzed uniformly based on account identifiers (ID).

[0003] Although the above method can analyze user habits with account ID and play a role in personalized recommendation, in real scenarios, account ID is used to identify a specific account, and generally one account can log in to multiple devices, so that users can share the account with natural persons for login and use. At this time, although different natural persons use the same account ID, the preferences or behavioral habits of different natural persons may be quite different. At this time, if the user behavior corresponding to the account ID in the existing relevant technology is still used as the basis for personalized recommendation, there will be a problem of not being able to make accurate personalized recommendations due to confused user behavior. Summary of the invention

[0004] In order to solve the problem that in the existing related technology, only the user behavior corresponding to the account is used as the basis for personalized recommendation, and accurate personalized recommendation cannot be performed due to confusion of user behavior, the present application provides a recommendation method, device, electronic device and storage medium.

[0005] In a first aspect, the present application provides a recommendation method, comprising:

[0006] Obtain account information and device information of user behavior;

[0007] Determine at least two device clusters corresponding to the account information;

[0008] Determine the device cluster to which the device information belongs as a target device cluster;

[0009] According to the historical behavior characteristics corresponding to the target device cluster and in combination with the user behavior, intention prediction is performed to generate recommendation information corresponding to the user behavior.

[0010] Optionally, determining at least two device clusters corresponding to the account information includes:

[0011] Determine at least two login devices corresponding to the account information;

[0012] Obtaining device behavior data from each of the login devices;

[0013] Perform feature extraction on the device behavior data to obtain device behavior features corresponding to the login device;

[0014] Each of the login devices is divided into clusters based on the device behavior characteristics to obtain at least two device clusters.

[0015] Optionally, clustering each of the login devices based on the device behavior characteristics to obtain at least two device clusters includes:

[0016] Comparing the difference between the behavioral characteristics of each two of the devices to obtain characteristic difference information;

[0017] When the characteristic difference information meets the preset cluster condition, cluster division is performed based on the login devices corresponding to the characteristic difference information to obtain the device cluster.

[0018] Optionally, the obtaining device behavior data in each of the login devices includes:

[0019] Obtaining behavior information, location information, and login time information from each of the login devices;

[0020] The device behavior data is generated based on the behavior information, the location information and the login time information.

[0021] Optionally, the predicting intention based on the historical behavior characteristics corresponding to the target device cluster and combining the user behavior to generate recommendation information corresponding to the user behavior includes:

[0022] Obtaining historical behavior characteristics corresponding to each target device in the target device cluster;

[0023] According to the historical behavior characteristics, the user behavior is predicted to obtain the recommendation information.

[0024] Optionally, the predicting the user behavior intention based on the historical behavior characteristics to obtain the recommendation information includes:

[0025] Based on the historical behavior characteristics, generating a target user portrait corresponding to the target device cluster;

[0026] The target user portrait is used to predict the user behavior intention and obtain the recommendation information.

[0027] Optionally, determining the device cluster to which the device information belongs as a target device cluster includes:

[0028] Extracting a device identifier from the device information;

[0029] The device cluster including the device identifier is determined as a target device cluster.

[0030] In a second aspect, a recommendation device is provided, comprising:

[0031] The acquisition module is used to obtain the account information and device information of user behavior;

[0032] A first determination module, configured to determine at least two device clusters corresponding to the account information;

[0033] A second determining module, configured to determine the device cluster to which the device information belongs as a target device cluster;

[0034] The intention prediction module is used to perform intention prediction based on the historical behavior characteristics corresponding to the target device cluster and the user behavior, and generate recommendation information corresponding to the user behavior.

[0035] In a third aspect, an electronic device is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0036] Memory, used to store computer programs;

[0037] A processor is used to implement the recommended method described in any one of the first aspects when executing a program stored in a memory.

[0038] According to a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the recommendation method according to any one of the first aspects is implemented.

[0039] The embodiment of the present application obtains the account information and device information of the user behavior to determine at least two device clusters corresponding to the account information, and determines the device cluster to which the device information belongs as the target device cluster, and then predicts the intention based on the historical behavior characteristics corresponding to the target device cluster and the user behavior, and generates recommendation information corresponding to the user behavior, so that the recommendation information can be generated specifically for the device cluster associated with the account information and the device information, thereby avoiding the problem in the existing related technology that only the user behavior corresponding to the account is used as the basis for personalized recommendation, and the inability to perform accurate personalized recommendation due to the confusion of user behavior. Recommendation information can be accurately and adaptively generated for each user, thereby improving the accuracy of recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0042] Figure 1 A flowchart of a recommended method provided in an embodiment of the present application;

[0043] Figure 2 A schematic diagram of a recommended device provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are 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 ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0046] As the functions of various applications and websites continue to increase, most operators will adopt a paid membership system to charge for different functional services. The charges are usually based on the account number, and one account can be logged in on multiple devices. At this time, most users will share the paid account, so that different natural persons log in to the same paid account on different devices; on the other hand, in order to improve user stickiness and user experience, the operator will adaptively push relevant content according to specific user behavior habits. As the number of account sharing increases, there will be anomalies in the process of personalized push of relevant recommended content. For example, account A is used by two natural persons. The behavior habit of natural person X is to frequently watch science and technology videos, while the behavior habit of natural person Y is to frequently watch history videos. According to the existing personalized recommendation strategy, all behaviors of account A will be collected for analysis and processing for recommendation. At this time, science and technology or history information will be pushed according to science and technology videos and history videos, that is, history information will be pushed to natural person X, or science and technology information will be pushed to natural person Y, so that natural person X and natural person Y will both receive push information that they are not interested in. That is, in the existing related technology, only the user behavior corresponding to the account is used as the basis for personalized recommendation, and there is a problem that accurate personalized recommendation cannot be made due to the confusion of user behavior.

[0047] In order to solve the problem that only the user behavior corresponding to the account is used as the basis for personalized recommendation in the existing related technology, and accurate personalized recommendation cannot be made due to the confusion of user behavior, the present application provides a recommendation method, apparatus, electronic device and storage medium, which obtains the account information and device information of the user behavior to determine at least two device clusters corresponding to the account information, and the device cluster to which the device information belongs is determined as the target device cluster, and then the intention is predicted based on the historical behavior characteristics corresponding to the target device cluster and the user behavior, and the recommendation information corresponding to the user behavior is generated, so that the recommendation information can be generated for the device cluster associated with the account information and the device information. This can avoid the problem that only the user behavior corresponding to the account is used as the basis for personalized recommendation in the existing related technology, and accurate personalized recommendation cannot be made due to the confusion of user behavior. The recommendation information can be accurately and adaptively generated for each user, thereby improving the accuracy of the recommendation.

[0048] Figure 1A flow chart of a method for determining a verification method provided in an embodiment of the present application. The method can be applied to one or more electronic devices such as a client and a server. In addition, the execution subject of the method can be hardware or software. When the above-mentioned execution subject is hardware, the execution subject can be one or more of the above-mentioned electronic devices. For example, a single electronic device can execute the method, or a plurality of electronic devices can cooperate with each other to execute the method. When the above-mentioned execution subject is software, the method can be implemented as a plurality of software or software modules, or as a single software or software module. No specific limitation is made herein.

[0049] like Figure 1 As shown, the present application discloses an embodiment, providing a recommendation method, which may specifically include the following steps:

[0050] Step S110: Obtaining account information and device information of user behavior;

[0051] Among them, user behavior can represent the behavioral data generated by the operations performed by the user in the client, such as clicking, watching videos, browsing content, etc., and when the user behavior is received, the account information and device information corresponding to the user behavior can be obtained, wherein the account information represents the information of the account logged in when the user behavior is implemented, such as account ID, account identification, etc., and the device information can represent the information of the device that implements the user behavior, that is, the device information can be used to indicate the specific device, and the device information can be the device ID, device model, device identification, etc., and the specific device can be a mobile device, computer and other devices.

[0052] Step S120: Determine at least two device clusters corresponding to the account information.

[0053] Specifically, after obtaining the account information, it can be determined that the account information corresponds to at least two preset device clusters. The device cluster can represent a cluster composed of devices that have logged in to the account information and have similar user behaviors. In addition, the device cluster can include at least one device identifier for marking a specific device.

[0054] It should be noted that the device cluster in this embodiment can be a cluster obtained by pre-dividing according to account information and device information. For example, account information A has four login devices, including login device A, login device B, login device C, and login device D. According to the pre-analysis and processing of user behavior, it is determined that the behaviors of login device A and login device C are relatively similar, both of which frequently watch science and technology videos, while the behaviors of login device B and login device D are relatively similar, both of which frequently watch historical videos. At this time, login device A and login device C, which may be used by the same natural person to log in, are divided into the same device cluster X, and login device B and login device D, which may be used by another natural person to log in, are divided into the same device cluster Y. In this way, through the combination of account information and device information, the user behavior of different natural persons using the same account information in different devices can be identified and distinguished, so that push information matching their behavior habits can be adaptively pushed to different natural persons according to the device cluster in the future. This can avoid the problem in the existing related technology that only the user behavior corresponding to the account is used as the basis for personalized recommendation, and the inability to perform accurate personalized recommendation due to chaotic user behavior, so as to achieve the effect of accurately and adaptively generating recommendation information for each natural person.

[0055] Step S130: Determine the device cluster to which the device information belongs as the target device cluster.

[0056] Specifically, after determining at least two device clusters corresponding to the account information, the device cluster to which the device information belongs can be determined, and the device cluster to which the device information belongs can be determined as the target device cluster. The device cluster to which the device information belongs can be determined by comparing the device information with the device identifier in each device cluster, and determining the device cluster containing the identifier corresponding to the device information as the target device cluster. Of course, other determination methods can also be used, and this embodiment does not make specific limitations on this.

[0057] Step S140: perform intention prediction based on the historical behavior characteristics corresponding to the target device cluster and in combination with the user behavior, and generate recommendation information corresponding to the user behavior.

[0058] Specifically, after determining the target device cluster, the historical behavior characteristics corresponding to the target device cluster can be obtained. The historical behavior information indicates the behavior characteristics of the devices marked by the target device cluster that are collected in advance, and based on the historical behavior characteristics, the intention is predicted in combination with the user behavior to generate recommendation information corresponding to the user behavior, that is, the historical behavior characteristics corresponding to the target device cluster and the user behavior are used to predict the intention together, so that the generated recommendation information can fit the behavior characteristics of each device in the target device cluster, and can also fit the current user behavior, so that all the behavior characteristics under one account information are no longer used as the basis for intent prediction, but the historical behavior characteristics matched by the device information corresponding to the current user behavior are used for intent prediction, thereby avoiding the problem of using all user behaviors corresponding to the account as the basis for personalized recommendations in the existing related technology, and the inability to perform accurate personalized recommendations due to confused user behaviors, so as to achieve the effect of accurately generating recommendation information adaptively for each natural person.

[0059] It can be seen that in this embodiment, the account information and device information of the user behavior are obtained to determine at least two device clusters corresponding to the account information, and the device cluster to which the device information belongs is determined as the target device cluster. Subsequently, the intention is predicted based on the historical behavior characteristics corresponding to the target device cluster and combined with the user behavior to generate recommendation information corresponding to the user behavior, so that the recommendation information can be generated for the device cluster associated with the account information and the device information. That is, the historical behavior characteristics that match the current user behavior are determined through the device information to predict the intention, thereby avoiding the problem of taking all user behaviors corresponding to the account as the basis for personalized recommendations in the existing related technology, and the inability to perform accurate personalized recommendations due to confused user behaviors. Recommendation information can be accurately and adaptively generated for each user, thereby improving the accuracy of recommendations.

[0060] In an optional implementation of the present application, step S120 determines at least two device clusters corresponding to the account information, which may specifically include the following sub-steps: determining at least two login devices corresponding to the account information; obtaining device behavior data in each login device; performing feature extraction on the device behavior data to obtain device behavior features corresponding to the login device; clustering each login device based on the device behavior features to obtain at least two device clusters.

[0061] In this embodiment, in the process of determining at least two device clusters corresponding to the account information, at least two login devices corresponding to the account information can be determined, and the login device identifies the device that logs in the account corresponding to the account information, so that the device behavior data in each login device can be obtained, and the device behavior data represents the behavior data generated in the device, which can be specifically the behavior data generated within a preset time, and the preset time can be 1 week, 10 days, etc.; then the device behavior data can be feature extracted to obtain the device behavior feature corresponding to the login device, and the feature extraction method can be a preset neural network model or algorithm. The device behavior feature can represent the characteristics of the user behavior in the login device, such as user preferences, user habits, etc., and then each login device can be clustered based on the device behavior feature to obtain at least two device clusters. Clustering is to divide the login devices with similar device behavior features into the same device cluster, so that at least two device clusters can be obtained, that is, the device cluster can represent the login devices with similar behavior features. It should be noted that the determination of at least two device clusters corresponding to the account information in this embodiment can be determined by executing the above steps in advance, or it can be determined by executing the above steps after obtaining the account information. This embodiment does not make specific restrictions on this.

[0062] In an optional implementation of the present application, each login device is clustered based on the device behavior characteristics to obtain at least two device clusters, which may specifically include the following sub-steps: comparing the differences between the behavior characteristics of each two devices to obtain characteristic difference information; when the characteristic difference information meets the preset cluster conditions, clustering is performed based on the login devices corresponding to the characteristic difference information to obtain a device cluster.

[0063] In this embodiment, during the cluster division process, the difference between the behavior characteristics of each two devices can be compared to obtain feature difference information, which represents the gap between the behavior characteristics of the two devices. The feature difference information can be a percentage difference or specific difference situation information, etc.; thereby determining whether the feature difference information meets the preset cluster conditions. The preset cluster conditions represent the conditions that can be divided into the same device cluster, for example, the percentage difference represented by the feature difference information is within 10 percent, etc., that is, when the feature difference information does not meet the preset cluster conditions, there is no need to divide the login device corresponding to the feature difference information into a cluster. When the feature difference information meets the preset cluster conditions, the login device corresponding to the feature difference information can be divided into the same cluster to obtain a device cluster, thereby playing the role of dividing login devices with similar device behavior characteristics into the same device cluster.

[0064] In an optional implementation of the present application, obtaining device behavior data in each login device may specifically include the following sub-steps: obtaining behavior information, location information and login time information in each login device; generating device behavior data based on the behavior information, location information and login time information.

[0065] In this embodiment, in the process of obtaining the device behavior data in each login device, the behavior information, location information and login time information in each login device can be obtained, wherein the behavior information can represent the information of the behavior implemented in the login device, such as click behavior, browsing behavior, interaction frequency behavior, etc., the location information can represent the information of the location of the login device, and the login time information represents the time when the login device logs into the account corresponding to the account information; then, based on the behavior information, location information and login time information, the device behavior data can be generated. Specifically, the behavior information, location information and login time information can be directly determined as the device behavior data, or the behavior information, location information and login time information can be processed to generate the device behavior data. The specific information processing method can be deduplication, screening out abnormal information, etc., and this embodiment does not make specific limitations on this.

[0066] In one example, the behavior information of the login device obtained may include movie viewing click behavior information, browsing behavior information, and interaction frequency information. The collection of movie viewing click behavior information may be recording the movie viewing click behavior information on the login device, such as the clicked elements, clicked locations, click times and frequencies, etc.; the acquisition of browsing behavior information may be analyzing the pages of movies browsed in the login device, the frequency of visits, the dwell time, the interaction paths, the depth of visits and other data to generate browsing behavior information; the interaction frequency information may be generated by analyzing the frequency of interaction between the user and the movie interface in the login device, such as scrolling, dragging, zooming in and out and other operation data; of course, the above is only an example for illustration, and adaptive adjustments may be made according to the scenario and needs during specific implementation, and this embodiment does not make specific limitations on this.

[0067] In addition, in the process of comparing the difference between the behavior characteristics of each two devices and obtaining the characteristic difference information, the behavior difference information between the behavior information behaviors can be compared, the time difference information between the login devices in different time periods can be compared, and the location difference information between the different login devices can be compared, and then the behavior difference information, time difference information and location difference information can be combined to perform difference statistics to generate the characteristic difference information; and when pre-determining the preset clustering conditions, the relevant conditions can also be adaptively configured according to the behavior information, location information and login time information, such as the preset clustering conditions can include behavior difference conditions, time difference conditions, location difference conditions, etc. The behavior difference conditions can be the difference degree of the behavior for the content field, the time difference conditions can be the difference degree of the login device time, and the location difference conditions can be the difference degree of the login device location, so that it can be judged from multiple angles whether the characteristic difference information meets the preset clustering conditions, and then the login devices corresponding to the characteristic difference information can be clustered more accurately. Of course, in the specific implementation, multiple rules can also be configured in the preset clustering conditions, for example, when two conditions are met in the behavior difference conditions, time difference conditions and location difference conditions, it can be determined that the characteristic difference information meets the preset clustering conditions, etc., and this embodiment does not make specific limitations on this.

[0068] In an optional implementation of the present application, step S140 predicts the intention of the user behavior based on the historical behavior characteristics corresponding to the target device cluster and the user behavior, and generates recommendation information corresponding to the user behavior. Specifically, it may include the following sub-steps: obtaining the historical behavior characteristics corresponding to each target device in the target device cluster; predicting the intention of the user behavior based on the historical behavior characteristics to obtain recommendation information.

[0069] In this embodiment, in the process of performing intention prediction, the historical behavior characteristics corresponding to each target device in the target device cluster can be obtained. The historical behavior characteristics represent the characteristics obtained after feature extraction of the historical behavior data corresponding to the target device. It can be the device behavior characteristics of the login device in the aforementioned embodiment, or it can be the characteristics obtained after feature extraction of the historical behavior data corresponding to the target device in real time. This embodiment does not make specific limitations on this. Then, based on all the historical behavior characteristics corresponding to the target device cluster, the user behavior can be predicted for intention to obtain recommendation information. The specific intention prediction method can be based on a preset prediction algorithm or a prediction neural network model to match the user behavior with the known historical behavior characteristics and dynamically generate recommended content. Of course, other intention prediction methods can also be used. This embodiment does not make specific limitations on this.

[0070] In an optional implementation of the present application, the user behavior intention is predicted based on historical behavior characteristics to obtain recommendation information, which may specifically include the following sub-steps: based on historical behavior characteristics, generating a target user portrait corresponding to the target device cluster; using the target user portrait, the user behavior intention is predicted to obtain recommendation information.

[0071] In this embodiment, in the process of intention prediction, since the historical behavior characteristics are the historical behavior characteristics corresponding to each target device in the target device cluster, a target user portrait corresponding to the target device cluster can be generated based on the historical behavior characteristics. The target user portrait represents the user portrait corresponding to the target device cluster, and one account information can correspond to at least two device clusters, that is, the account information can correspond to multiple user portraits, so that when the account is shared, it can be determined through the target device cluster to draw a corresponding target user portrait for each user using the account information; and then the target user portrait is used to predict the user behavior and obtain recommendation information, that is, to adaptively predict the intention for each user using the account information, and obtain recommendation information that meets the current user, thereby avoiding the problem of using all user behaviors corresponding to the account as the basis for personalized recommendation in the existing related technology, and the inability to perform accurate personalized recommendation due to confusion in user behavior, and can accurately generate recommendation information for the user corresponding to each device adaptively, thereby providing each user with adapted personalized content push welfare activities on different devices used by him, and the specific recommendation information can be video, advertising recommendation and other content.

[0072] It should be noted that the device behavior of the login device or the historical behavior of the target device, as well as related expressions, in this application, all refer to the behavior of the user interacting with the client on the device corresponding to the account information of the login account, and do not refer to the hardware behavior of the device.

[0073] In an optional implementation of the present application, step S130 determines the device cluster to which the device information belongs as the target device cluster, which may specifically include the following sub-steps: extracting the device identifier from the device information; and determining the device cluster containing the device identifier as the target device cluster.

[0074] In this embodiment, in the process of determining the device cluster to which the device information belongs as the target device cluster, a device identifier can be extracted from the device information. The device identifier represents the device used to mark the device that implements the current user behavior, and the device cluster can include identifiers corresponding to each device belonging to a cluster. The device identifier can be compared with the identifier in the device cluster to determine whether there is an identifier matching the device identifier in each device cluster, and the device cluster with an identifier matching the device identifier can be determined as the target device cluster.

[0075] like Figure 2 As shown, the present application also discloses an embodiment, providing a recommendation device, which may include the following modules:

[0076] The acquisition module 210 is used to obtain the account information and device information of the user behavior;

[0077] A first determination module 220, configured to determine at least two device clusters corresponding to the account information;

[0078] A second determination module 230, configured to determine the device cluster to which the device information belongs as a target device cluster;

[0079] The intention prediction module 240 is used to perform intention prediction based on the historical behavior characteristics corresponding to the target device cluster and in combination with the user behavior, and generate recommendation information corresponding to the user behavior.

[0080] In an optional implementation of the present application, the first determining module 220 may include the following units:

[0081] A first determining unit, configured to determine at least two login devices corresponding to the account information;

[0082] A first acquisition unit, used to acquire device behavior data from each of the login devices;

[0083] A first feature extraction unit, configured to extract features from the device behavior data to obtain device behavior features corresponding to the login device;

[0084] The cluster division unit is used to divide each of the login devices into clusters based on the device behavior characteristics to obtain at least two device clusters.

[0085] In an optional implementation of the present application, the cluster division unit may include the following subunits:

[0086] A comparison subunit, used for comparing the difference between the behavioral features of each two of the devices to obtain feature difference information;

[0087] The cluster division subunit is used to perform cluster division based on the login devices corresponding to the feature difference information to obtain the device cluster when the feature difference information meets the preset cluster condition.

[0088] In an optional implementation of the present application, the first acquisition unit may include the following subunits:

[0089] An acquisition subunit, used to acquire behavior information, location information and login time information from each of the login devices;

[0090] The first generating subunit is used to generate the device behavior data based on the behavior information, the location information and the login time information.

[0091] In an optional implementation of the present application, the intention prediction module 240 may include the following units:

[0092] A second acquisition unit, configured to acquire a historical behavior feature corresponding to each target device in the target device cluster;

[0093] The intention prediction unit is used to predict the intention of the user behavior based on the historical behavior characteristics to obtain the recommendation information.

[0094] In an optional implementation of the present application, the intention prediction unit may include the following subunits:

[0095] A second generating subunit, configured to generate a target user portrait corresponding to the target device cluster based on the historical behavior characteristics;

[0096] The intention prediction subunit is used to predict the intention of the user behavior by using the target user portrait to obtain the recommendation information.

[0097] In an optional implementation of the present application, the second determining module 230 may include the following units:

[0098] An extraction unit, used to extract a device identifier from the device information;

[0099] The second determining unit is configured to determine the device cluster including the device identifier as a target device cluster.

[0100] The implementation process of the functions and effects of each module in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, which will not be repeated here.

[0101] like Figure 3 As shown, an embodiment of the present application provides an electronic device, including a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340;

[0102] Memory 330, for storing computer programs;

[0103] In one embodiment of the present application, the processor 310, when executing the program stored in the memory 330, implements the recommendation method provided by any of the aforementioned method embodiments, by obtaining the account information and device information of the user behavior to determine at least two device clusters corresponding to the account information, and the device cluster to which the device information belongs is determined as the target device cluster, and then the intention is predicted based on the historical behavior characteristics corresponding to the target device cluster and the user behavior, and recommendation information corresponding to the user behavior is generated, so that the recommendation information can be generated for the device cluster associated with the account information and the device information, that is, the historical behavior characteristics that match the current user behavior are determined through the device information to predict the intention, thereby avoiding the problem of using all user behaviors corresponding to the account as the basis for personalized recommendations in the existing related technology, and the inability to perform accurate personalized recommendations due to confusion in user behaviors, and can accurately and adaptively generate recommendation information for each user, thereby improving the accuracy of recommendations.

[0104] The embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a recommendation method as provided in any of the aforementioned method embodiments, by obtaining account information and device information of user behavior to determine at least two device clusters corresponding to the account information, and determining the device cluster to which the device information belongs as a target device cluster, and then generating recommendation information corresponding to the user behavior based on historical behavior characteristics corresponding to the target device cluster and combining the user behavior to perform intention prediction, thereby generating recommendation information for the device cluster associated with the account information and the device information, that is, determining the historical behavior characteristics that match the current user behavior through the device information to perform intention prediction, thereby avoiding the problem in the existing related technology of using all user behaviors corresponding to the account as the basis for personalized recommendation, and being unable to perform accurate personalized recommendation due to confusion in user behavior, and being able to accurately and adaptively generate recommendation information for each user, thereby improving the accuracy of recommendation.

[0105] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0106] The above describes specific embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0107] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may 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 the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A recommendation method, characterized in that: include: Obtain account information and device information of user behavior; Determine at least two device clusters corresponding to the account information; Determine the device cluster to which the device information belongs as a target device cluster; According to the historical behavior characteristics corresponding to the target device cluster and in combination with the user behavior, intention prediction is performed to generate recommendation information corresponding to the user behavior.

2. The recommendation method according to claim 1, characterized in that: Determining at least two device clusters corresponding to the account information includes: Determine at least two login devices corresponding to the account information; Obtaining device behavior data from each of the login devices; Perform feature extraction on the device behavior data to obtain device behavior features corresponding to the login device; Each of the login devices is divided into clusters based on the device behavior characteristics to obtain at least two device clusters.

3. The recommendation method according to claim 2, characterized in that: The clustering of each of the login devices based on the device behavior characteristics to obtain at least two device clusters includes: Comparing the difference between the behavioral characteristics of each two of the devices to obtain characteristic difference information; When the characteristic difference information meets the preset cluster condition, cluster division is performed based on the login devices corresponding to the characteristic difference information to obtain the device cluster.

4. The recommendation method according to claim 2, characterized in that: The obtaining of device behavior data in each of the login devices includes: Obtaining behavior information, location information, and login time information from each of the login devices; The device behavior data is generated based on the behavior information, the location information and the login time information.

5. The recommendation method according to claim 1, characterized in that: The intention prediction is performed based on the historical behavior characteristics corresponding to the target device cluster and in combination with the user behavior to generate recommendation information corresponding to the user behavior, including: Obtaining historical behavior characteristics corresponding to each target device in the target device cluster; According to the historical behavior characteristics, the user behavior is predicted to obtain the recommendation information.

6. The recommendation method according to claim 5, characterized in that: The step of predicting the user behavior intention based on the historical behavior characteristics to obtain the recommendation information includes: Based on the historical behavior characteristics, generating a target user portrait corresponding to the target device cluster; The target user portrait is used to predict the user behavior intention and obtain the recommendation information.

7. The recommendation method according to any one of claims 1 to 6, characterized in that: The step of determining the device cluster to which the device information belongs as a target device cluster includes: Extracting a device identifier from the device information; The device cluster including the device identifier is determined as a target device cluster.

8. A recommendation device, characterized in that: include: The acquisition module is used to obtain the account information and device information of user behavior; A first determination module, configured to determine at least two device clusters corresponding to the account information; A second determining module, configured to determine the device cluster to which the device information belongs as a target device cluster; The intention prediction module is used to perform intention prediction based on the historical behavior characteristics corresponding to the target device cluster and the user behavior, and generate recommendation information corresponding to the user behavior.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, used to implement the recommendation method described in any one of claims 1 to 7 when executing a program stored in a memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the recommendation method according to any one of claims 1 to 7 is implemented.

Images