Recommendation method, device, equipment and storage medium based on user behavior

Abstract

The present invention relates to the field of data analysis, and discloses a recommendation method, device, equipment and storage medium based on user behavior. The method includes: obtaining the operation behavior data of users on various application platforms through auxiliary function events; marking the user with a behavior tag according to the operation behavior data, and obtaining the weight corresponding to the behavior tag according to the frequency of the behavior tag triggering; selecting the behavior tag according to the weight corresponding to the behavior tag; extracting the user account information, forming the user's full network tag according to the selected behavior tag and the user account information, and recommending relevant content to the user in the next application platform browsing according to the user's full network tag. In an embodiment of the present invention, the user's information on various application platforms can be integrated to effectively prevent the fault of user information, thereby improving the accuracy of the recommendation.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to a recommendation method, device, equipment and storage medium based on user behavior. Background Art

[0002] With the widespread use of smart devices, most people's information is currently stored on the Internet or smart devices. However, people's preferences vary significantly. In order to help users access content they are interested in more quickly and conveniently and achieve personalized recommendations, analyzing user behavior habits can help applications and services provide more customized recommendation services, such as music, videos, news, etc. By studying users' daily activities, smart assistants can predict users' needs and make corresponding preparations in advance, such as automatically adjusting alarms, reminding meetings, etc.

[0003] Current data analysis and behavior detection technologies are already quite mature, mainly relying on big data analysis, machine learning and artificial intelligence algorithms. These technologies can monitor user behavior in real time, analyze usage habits, and provide personalized recommendations, optimize user experience and ensure device security. However, user behavior tracking on the market now is mainly based on a single app and browser. When switching apps or the browser switches to another primary domain name, user information will be lost, resulting in a gap in user information. Summary of the invention

[0004] The main purpose of the present invention is to solve the technical problem of user information gap.

[0005] A first aspect of the present invention provides a recommendation method based on user behavior, the recommendation method based on user behavior comprising:

[0006] Acquire the user's operation behavior data on each application platform through auxiliary function events, wherein the auxiliary function events are used to receive the callback content of the application platform interface information to obtain the current interface information;

[0007] According to the operation behavior data, a behavior tag is added to the user, and according to the triggering frequency of the behavior tag, a weight corresponding to the behavior tag is obtained;

[0008] Selecting the behavior tag according to the weight corresponding to the behavior tag;

[0009] Extract user account information, form a user-wide network tag based on the selected behavior tag and the user account information, and recommend relevant content to the user in the next application platform browsing based on the user-wide network tag.

[0010] Optionally, in a first implementation of the first aspect of the present invention, extracting user account information, forming a user network-wide tag according to the selected behavior tag and the user account information, and recommending relevant content to the user in the next application platform browsing according to the user network-wide tag includes:

[0011] Obtaining HTTP data packets of the user on each application platform;

[0012] Viewing the HTTP data packet through a filter to obtain user account information;

[0013] According to the selected behavior tag and the user account information, a user network-wide tag is formed;

[0014] Based on the user's full network tags, relevant content is recommended to the user the next time the user browses the application platform.

[0015] Optionally, in a second implementation of the first aspect of the present invention, forming a user network-wide tag according to the selected behavior tag and the user account information includes:

[0016] Deduplicating the user account information, and accumulating the frequency of the deduplicated user account information;

[0017] Sorting the user account information after frequency accumulation, and taking a preset number of user account information as the user identification of the device;

[0018] A user network-wide label is formed based on the selected behavior label and the user identifier.

[0019] Optionally, in a third implementation of the first aspect of the present invention, the operation behavior data includes a list of accessed contents, access duration, and behavior event type for each application platform;

[0020] The step of labeling the user with a behavior tag according to the operation behavior data, and obtaining a weight corresponding to the behavior tag according to a triggering frequency of the behavior tag includes:

[0021] According to the access content list, the access content duration and the behavior event type, a behavior tag is added to the user;

[0022] The weight corresponding to the behavior tag is obtained according to the frequency of triggering the behavior tag and the duration of accessing the content.

[0023] Optionally, in a fourth implementation of the first aspect of the present invention, the step of labeling the user with a behavior tag according to the access content list, the access content duration, and the behavior event type includes:

[0024] According to the access content list, obtaining keywords in the access content list by a text analysis method;

[0025] According to the keywords, the subject of the accessed content is obtained;

[0026] A behavior label is assigned to the user according to the subject of the accessed content, the duration of the accessed content, and the type of the behavior event.

[0027] Optionally, in a fifth implementation of the first aspect of the present invention, before obtaining the user's operation behavior data on each application platform through auxiliary function events, the method further includes:

[0028] Registering an accessibility service to distribute interaction information of the current device interface content to applications registered with the accessibility service;

[0029] The callback of the auxiliary function event is specified so that when the interface information of each application platform changes, the application program registered with the auxiliary function service can receive the corresponding application platform message.

[0030] Optionally, in a sixth implementation manner of the first aspect of the present invention, before extracting the user account information and forming a user network-wide tag according to the selected behavior tag and the user account information, the method further includes:

[0031] Detecting whether the operation behavior data has changed;

[0032] If the operation behavior data changes, the behavior label is replaced;

[0033] According to the replaced behavior label, the weight corresponding to the behavior label is updated.

[0034] A second aspect of the present invention provides a user behavior-based recommendation device, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via lines; the at least one processor calls the instructions in the memory so that the user behavior-based recommendation device executes the above-mentioned user behavior-based recommendation method.

[0035] A third aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned recommendation method based on user behavior.

[0036] In an embodiment of the present invention, the user's operational behavior data on each application platform is obtained through auxiliary function events, and the auxiliary function events are used to receive the callback content of the application platform interface information to obtain the current interface information; a behavioral label is added to the user based on the operational behavior data, and the weight corresponding to the behavioral label is obtained based on the frequency of triggering the behavioral label; the behavioral label is selected based on the weight corresponding to the behavioral label; the user account information is extracted, and a user full-network label is formed based on the selected behavioral label and the user account information, and relevant content is recommended to the user in the next application platform browsing based on the user full-network label. In the present invention, the user's operational behavior data on each application platform is obtained through auxiliary function events, and then the user is labeled based on the operational behavior data to recommend relevant content to the user in the next application platform browsing, which can integrate the user's information on each application platform, effectively prevent the discontinuity of user information, and thus improve the accuracy of the recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of an embodiment of a recommendation method based on user behavior in an embodiment of the present invention;

[0038] Figure 2 A schematic diagram of an embodiment of a recommendation device based on user behavior in an embodiment of the present invention;

[0039] Figure 3 The figure is a schematic diagram of an embodiment of a recommendation device based on user behavior in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] Embodiments of the present invention provide a method, apparatus, device and storage medium for recommendation based on user behavior.

[0041] The embodiments disclosed in the present invention will be described in more detail below with reference to the accompanying drawings. Although certain embodiments disclosed in the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments disclosed in the present invention are only for exemplary purposes and are not intended to limit the scope of protection disclosed in the present invention.

[0042] In the description of the embodiments disclosed in the present invention, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0043] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the recommendation method based on user behavior in an embodiment of the present invention includes:

[0044] S100, obtaining the user's operation behavior data on each application platform through auxiliary function events.

[0045] In this embodiment, an Android application is written, and auxiliary functions are set in the manifest file of the application, and the user's operation behavior data on each application platform is obtained from the auxiliary function event, wherein the auxiliary function event is used to receive the callback content of the application platform interface information to obtain the current interface information, specifically, by defining an interface and declaring an instance of the interface in the class. When an auxiliary function event occurs, the system will call the callback method in the interface to handle the event, thereby obtaining the current interface information. The advantage of this method is that the interface and the implementation are separated, and the implementation class can be flexibly replaced, thereby achieving the purpose of interface unification. Then, the operation behavior data is analyzed from the current interface information through the user behavior analysis method, wherein the user behavior analysis method includes behavior event analysis, page click analysis, descriptive statistical analysis, funnel analysis, etc. After obtaining the operation behavior data, it is recorded in the database for storage.

[0046] S200, according to the operation behavior data, a behavior tag is marked for the user, and according to the frequency of triggering of the behavior tag, a weight corresponding to the behavior tag is obtained.

[0047] In this embodiment, it is first necessary to collect the user's interface information data, which includes various operation behaviors and interface display information performed by the user in various application platforms, wherein the operation behaviors include opening the application, closing the application, clicking on the content of the picture, enlarging the picture / webpage, clicking on the button, keyboard input information content, etc. By analyzing these interface information data, the user's behavior patterns in different interfaces can be understood. For example, the user's frequently visited interface and the button with a high click rate can be identified. Then, based on the analysis results, a set of label systems is defined. This system can include structured labels, semi-structured labels and unstructured labels. The label system should be able to cover all aspects of user behavior, such as interface functions, user preferences, operation habits, etc., and then use AI technology to automatically identify the key elements in the interface content and label these elements according to the defined label system. In addition, in order to ensure the accuracy and effectiveness of the label, user testing can be carried out. Through user testing, it can be verified whether the label can correctly convey information and be optimized based on feedback.

[0048] S300: Select a behavior tag according to the weight corresponding to the behavior tag.

[0049] In this embodiment, the frequency of behavior tag triggering is counted to obtain the weight corresponding to the behavior tag, and the behavior tags are sorted according to the weight. Based on the sorting, the behavior tags are selected, for example, the first 10 behavior tags sorted by weight are retained.

[0050] S400, extracting user account information, forming a user network-wide tag based on the selected behavior tag and user account information, and recommending relevant content to the user in the next application platform browsing based on the user network-wide tag.

[0051] In this embodiment, user account information is extracted to achieve centralized and unified management of the user account life cycle, and a synchronization mechanism with each application platform is established. The tags selected by users on different platforms are integrated, and a comprehensive user full-network tag is formed based on the selected behavioral tags and user account information, which can be achieved through cross-platform data synchronization and tag matching. Then, based on the user's full-network tag, a recommendation system is constructed, wherein the recommendation system can adopt a content-based recommendation method to recommend similar subject matter to users based on the user's historical behavior and relevant information of the subject matter. In addition, collaborative filtering, knowledge-based recommendation and other methods can be combined to improve the accuracy and diversity of recommendations. When the user browses the application platform next time, relevant content is recommended based on its full-network tag, and the recommended content should be as consistent as possible with the user's interests and behavioral characteristics. The recommendation method of the present invention can be applied to smart phone devices, and tablet devices can also be compatible.

[0052] In an optional implementation of the first aspect of the present invention, extracting user account information, forming a user network-wide tag based on the selected behavior tag and the user account information, and recommending relevant content to the user in the next application platform browsing based on the user network-wide tag includes:

[0053] Obtain the user's HTTP data packets on each application platform; view the HTTP data packets through filters to obtain user account information; form user-wide network tags based on selected behavior tags and user account information; based on the user's full network tags, recommend relevant content to the user the next time he browses the application platform.

[0054] In this embodiment, each application platform authorizes the device so that the device can extract the user's account information (mobile phone / email) on each application platform. Optionally, HTTP data packets can be captured through Wireshark to analyze the login information in the POST request. For example, when logging into an HTTP website, fill in the user name and password and submit, then check the Wireshark interface, you can see a POST login data packet, which contains the input account number. Specifically, after starting Wireshark, select the correct network interface for data packet capture. In order to capture only HTTP data packets related to the local host, you can use a capture filter. For example, it can be set to only capture data packets whose source IP is the local host; check the filtered HTTP data packets and find the field containing the user account information. Since the HTTP protocol transmits plain text data, this information will be directly displayed in the data packet. Then, based on the selected behavior label and the extracted user account information, a user full network label is formed.

[0055] In an optional implementation of the first aspect of the present invention, forming a user network-wide tag according to the selected behavior tag and user account information includes:

[0056] De-duplicate the user account information, and accumulate the frequency of the de-duplicated user account information; sort the user account information after the frequency accumulation, and take a preset number of user account information as the user identification of the device; form a user full network label based on the selected behavior label and user identification.

[0057] In this embodiment, information is distributed to each application platform through a written Android application to extract data from the account and password interfaces of all application platforms, and record the user account information (mobile phone / email) of each application platform logged in on the device, and store the extracted user account information in a list or database. Then, deduplication and frequency accumulation are performed, and the user account information with the highest frequency of a preset number is taken as the user id of this device, where the preset number can be 3. Optionally, the account information can be deduplicated using Python's set data structure. The frequency of the deduplicated account information is accumulated, and a dictionary (dict) can be used to record the number of logins for each account. Then, the dictionary after frequency accumulation is sorted, and the three account information with the highest frequency are taken, and these three account information are used as the user id of this device.

[0058] In an optional implementation of the first aspect of the present invention, the operation behavior data includes a list of accessed content, access duration, and behavior event type for each application platform; based on the operation behavior data, a behavior tag is added to the user, and based on the frequency of triggering the behavior tag, a weight corresponding to the behavior tag is obtained, including:

[0059] Based on the list of accessed content, the duration of access to the content, and the type of behavioral event, users are labeled with behavioral tags; based on the frequency of behavioral tag triggering and the duration of access to the content, the corresponding weight of the behavioral tag is obtained.

[0060] In this embodiment, the operation behavior data includes the access content list, access content duration and behavior event type of each application platform, wherein the access content list includes the content display list of existing application platform products, the access content duration includes the time the user stays on the application platform interface, and the behavior event type includes clicking the screen, the content change of the screen, the lock screen, the interface change of the application, the sliding event, etc. The weight calculation of the user behavior tag usually involves multiple factors, including the behavior event type, the frequency of the behavior tag triggering, etc., wherein the weight of the behavior event type is generally determined by the operator or the business, and different behaviors (such as browsing, searching, collecting, placing orders, purchasing, etc.) have different importance to users. The frequency of the behavior tag triggering reflects the degree of interest in the behavior. The higher the frequency, the higher the weight. Optionally, the weight of the user behavior tag can take into account time decay. The user's behavior will decay over time, that is, recent behavior is more representative than long-term behavior, and the recent behavior has a higher weight. Therefore, according to the access content list, the access content duration and the behavior event type, the user can be marked with a behavior tag, and the weight corresponding to the behavior tag can be obtained according to the frequency of the behavior tag triggering and the access content duration.

[0061] In an optional implementation of the first aspect of the present invention, labeling a user's behavior according to the access content list, the access content duration, and the behavior event type includes:

[0062] Based on the visited content list, the keywords in the visited content list are obtained through text analysis methods; based on the appearance of keywords, the theme of the visited content is obtained; based on the theme of the visited content, the duration of the visited content and the type of behavioral event, the user is labeled with a behavioral label.

[0063] In this embodiment, the access content list of each application platform is first obtained, and the word frequency statistics method is used to extract the keywords in the access content list. This method focuses on the frequency of occurrence of each word in the text. By counting and sorting these frequencies, the importance and usage frequency of each word in the text can be understood. Then, the theme in the text can be identified through the word frequency statistics and topic modeling methods, among which topic modeling can reveal the implicit theme in the text and identify the theme by analyzing the words that often appear together in the document collection. According to the theme of the access content, the access duration and the type of behavioral event, the user can be labeled with a behavioral label. For example, if the user mainly browses articles related to "water quality", he can be marked as a user interested in the field of "water quality".

[0064] In an optional implementation of the first aspect of the present invention, before obtaining the user's operation behavior data on each application platform through auxiliary function events, the method further includes:

[0065] Register the accessibility service to distribute the interactive information of the current device interface content to the application registered for the accessibility service; specify the callback of the accessibility event so that when the interface information of each application platform changes, the application registered for the accessibility service can receive the corresponding application platform message.

[0066] In this embodiment, an Android application is written, and the accessibility function is set in the manifest file of the application. In the accessibility function configuration, the callback content that the application needs to receive system interface information is set. The accessibility function is a service provided by the Android system. It can distribute interactive information of the current device interface content (such as a button clicked by the user, content information viewed by the user) and the ability to create a top-level view to applications that register for the service. For example, when a mobile phone user opens an app or the content of the screen changes, the system will distribute corresponding messages to all applications that have registered the accessibility function, so that the applications can know the changes in the current interface.

[0067] In an optional implementation of the first aspect of the present invention, before extracting the user account information and forming the user network-wide tag according to the selected behavior tag and the user account information, the method further includes:

[0068] Detect whether the operation behavior data has changed; if the operation behavior data has changed, replace the behavior label; based on the replaced behavior label, update the weight corresponding to the behavior label.

[0069] In this embodiment, when the operation behavior data changes, the behavior label needs to be updated first. This process can be achieved through a variety of algorithms, such as a dynamic weighted hybrid recommendation algorithm based on user behavior characteristics, which quantifies user behavior characteristics and introduces them into user similarity calculations, and adaptively connects multiple recommendation algorithms for prediction. When updating the weights corresponding to the behavior labels, a multiplication weight update method can be used. This method initially assigns the same weight to each behavior label and adjusts the weight through each prediction. User labels are continuously updated as user behavior changes, which can make the recommended related content more quickly close to the user's recent preferences.

[0070] See also Figure 2 According to a second aspect of the present invention, a recommendation device based on user behavior is provided, and the recommendation device based on user behavior comprises:

[0071] The behavior data acquisition module 10 is used to acquire the user's operation behavior data on each application platform through auxiliary function events. The auxiliary function events are used to receive the callback content of the application platform interface information to obtain the current interface information;

[0072] The behavior tag acquisition module 20 is used to tag the user with a behavior tag based on the operation behavior data, and obtain the weight corresponding to the behavior tag based on the frequency of the behavior tag being triggered;

[0073] A behavior tag selection module 30, used to select a behavior tag according to a weight corresponding to the behavior tag;

[0074] The content recommendation module 40 is used to extract user account information, form a user network-wide tag based on the selected behavior tag and user account information, and recommend relevant content to the user in the next application platform browsing based on the user network-wide tag.

[0075] In an optional implementation of the second aspect of the present invention, the content recommendation module 40 is also used to obtain HTTP data packets of users on various application platforms; view HTTP data packets through filters to obtain user account information; form user full-network tags based on selected behavior tags and user account information; and recommend relevant content to users when they browse the application platform next time based on the user full-network tags.

[0076] In an optional implementation of the second aspect of the present invention, the content recommendation module 40 is also used to deduplicate user account information and accumulate the frequency of the deduplicated user account information; sort the user account information after frequency accumulation, and take a preset number of user account information as the user identification of the device; and form a user full network label based on the selected behavior label and user identification.

[0077] In an optional implementation of the second aspect of the present invention, the behavior tag acquisition module 20 is also used to tag the user with a behavior tag based on the access content list, the duration of content access and the type of behavior event; and obtain the weight corresponding to the behavior tag based on the frequency of behavior tag triggering and the duration of content access.

[0078] In an optional implementation of the second aspect of the present invention, the behavior label acquisition module 20 is also used to obtain keywords in the accessed content list through a text analysis method based on the accessed content list; obtain the subject of the accessed content based on the keywords; and label the user with a behavior label based on the subject of the accessed content, the duration of the accessed content, and the type of behavior event.

[0079] In an optional implementation of the second aspect of the present invention, the user behavior-based recommendation device also includes an accessibility registration module, which is used to register accessibility services to distribute interactive information of the current device interface content to applications that register accessibility services; and specify callbacks for accessibility events so that when the interface information of each application platform changes, the applications that register accessibility services can receive corresponding application platform messages.

[0080] In an optional implementation of the second aspect of the present invention, the recommendation device based on user behavior also includes a behavior data detection module for detecting whether the operation behavior data has changed; if the operation behavior data has changed, the behavior label is replaced; and according to the replaced behavior label, the weight corresponding to the behavior label is updated.

[0081] Figure 3 is a schematic diagram of the structure of a user behavior-based recommendation device provided by an embodiment of the present invention. The user behavior-based recommendation device 500 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 510 (for example, one or more processors) and a memory 520, and one or more storage media 530 (for example, one or more mass storage devices) storing application programs 533 or data 532. Among them, the memory 520 and the storage medium 530 may be temporary storage or permanent storage. The program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the user behavior-based recommendation device 500. Furthermore, the processor 510 may be configured to communicate with the storage medium 530 to execute a series of instruction operations in the storage medium 530 on the user behavior-based recommendation device 500.

[0082] The user behavior-based recommendation device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Figure 3 The illustrated structure of the user behavior-based recommendation device does not constitute a limitation on the user behavior-based recommendation device, and may include more or less components than illustrated, or a combination of certain components, or a different arrangement of components.

[0083] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the user behavior-based recommendation method.

[0084] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0085] In addition, although each operation is described in a specific order, this should be understood as requiring such operation to be performed in the specific order shown or in a sequential order, or requiring that all illustrated operations should be performed to obtain desired results. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single implementation in combination. On the contrary, the various features described in the context of a single implementation can also be implemented in multiple implementations individually or in any suitable sub-combination mode.

[0086] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. A recommendation method based on user behavior, characterized in that: The recommendation method based on user behavior includes: Acquire the user's operation behavior data on each application platform through auxiliary function events, wherein the auxiliary function events are used to receive the callback content of the application platform interface information to obtain the current interface information; According to the operation behavior data, a behavior tag is added to the user, and according to the triggering frequency of the behavior tag, a weight corresponding to the behavior tag is obtained; Selecting the behavior tag according to the weight corresponding to the behavior tag; Extract user account information, form a user-wide network tag based on the selected behavior tag and the user account information, and recommend relevant content to the user in the next application platform browsing based on the user-wide network tag.

2. The user behavior-based recommendation method according to claim 1, characterized in that: The extracting of user account information, forming a user network-wide tag according to the selected behavior tag and the user account information, and recommending relevant content to the user in the next application platform browsing according to the user network-wide tag includes: Obtaining HTTP data packets of the user on each application platform; Viewing the HTTP data packet through a filter to obtain user account information; According to the selected behavior tag and the user account information, a user network-wide tag is formed; Based on the user's full network tags, relevant content is recommended to the user the next time the user browses the application platform.

3. The user behavior-based recommendation method according to claim 2, characterized in that: The forming of a user network-wide tag according to the selected behavior tag and the user account information includes: Deduplicating the user account information, and accumulating the frequency of the deduplicated user account information; Sorting the user account information after frequency accumulation, and taking a preset number of user account information as the user identification of the device; A user network-wide label is formed based on the selected behavior label and the user identifier.

4. The user behavior-based recommendation method according to claim 1, characterized in that: The operation behavior data includes a list of accessed content, access duration, and behavior event type for each application platform; The step of labeling the user with a behavior tag according to the operation behavior data, and obtaining a weight corresponding to the behavior tag according to a triggering frequency of the behavior tag includes: According to the access content list, the access content duration and the behavior event type, a behavior tag is added to the user; The weight corresponding to the behavior tag is obtained according to the frequency of triggering the behavior tag and the duration of accessing the content.

5. The user behavior-based recommendation method according to claim 4, characterized in that: The step of labeling the user with a behavior tag according to the access content list, the access content duration, and the behavior event type includes: According to the access content list, obtaining keywords in the access content list by a text analysis method; According to the keywords, the subject of the accessed content is obtained; A behavior label is assigned to the user according to the subject of the accessed content, the duration of the accessed content, and the type of the behavior event.

6. The user behavior-based recommendation method according to claim 1, characterized in that: Before obtaining the user's operation behavior data on each application platform through auxiliary function events, the method further includes: Registering an accessibility service to distribute interaction information of the current device interface content to applications registered with the accessibility service; The callback of the auxiliary function event is specified so that when the interface information of each application platform changes, the application program registered with the auxiliary function service can receive the corresponding application platform message.

7. The user behavior-based recommendation method according to claim 1, characterized in that: The extracting of the user account information, before forming a user network-wide tag according to the selected behavior tag and the user account information, further includes: Detecting whether the operation behavior data has changed; If the operation behavior data changes, the behavior label is replaced; According to the replaced behavior label, the weight corresponding to the behavior label is updated.

8. A recommendation device based on user behavior, characterized in that: The recommendation device based on user behavior includes: A behavior data acquisition module is used to acquire the user's operation behavior data on each application platform through auxiliary function events. The auxiliary function events are used to receive the callback content of the application platform interface information to obtain the current interface information; A behavior tag acquisition module, used to tag the user with a behavior tag according to the operation behavior data, and obtain a weight corresponding to the behavior tag according to the frequency of triggering the behavior tag; A behavior label selection module, used for selecting the behavior label according to the weight corresponding to the behavior label; The content recommendation module is used to extract user account information, form a user network-wide tag based on the selected behavior tag and the user account information, and recommend relevant content to the user in the next application platform browsing based on the user network-wide tag.

9. A recommendation device based on user behavior, characterized in that: The user behavior-based recommendation device includes: a memory and at least one processor, the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor calls the instructions in the memory to enable the user behavior-based recommendation device to perform the user behavior-based recommendation method as described in any one of claims 1-7.

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 user behavior-based recommendation method according to any one of claims 1 to 7 is implemented.

Images