Service recommendation method and device based on user behaviors, equipment and storage medium

Abstract

The invention belongs to the technical field of artificial intelligence, is applied to the field of financial science and technology and the field of medical health, and discloses a user behavior-based service recommendation method, device and equipment and a storage medium, and the user behavior-based service recommendation method comprises the following steps: obtaining user behavior information of a target user for a target service; analyzing the user behavior information, and generating real-time search data of the user; obtaining user feature information of a target user; inputting the real-time search data and the user feature information into a vectorization recall model to obtain an alternative service set; according to a preset user preference degree parameter, evaluating scores of services in the alternative service set by the target user, and sorting the scores to obtain a sorted alternative service sequence set; and taking the service which is ranked before the preset digit in the alternative service sequence set as the optimal recommendation service of the target user. The technical problem that in the prior art, the service product recommended to the user does not conform to the real demand of the user is solved.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology and is applied to the fields of financial technology and medical health. In particular, it relates to a business recommendation method, device, equipment and storage medium based on user behavior. Background Art

[0002] In today's society, user demand determines the sales rate of business products and the profits of business companies. Therefore, pushing business products that truly meet users' needs to users has gradually become the direction pursued by various business companies. For example, in the insurance business, such as property insurance, users' own financial reserves are limited, and the service life of their vehicles is relatively short. If long-term insurance is recommended to users without considering their personal conditions and essential needs, and long-term insurance is expensive, it obviously does not meet the real needs of users, and users will not purchase such business products that do not meet their needs. For example, in the life insurance field, when users are in their youth, the incidence rate is obviously low at this stage, and the user's personal funds are obviously insufficient at this stage. If critical illness insurance is purchased (critical illness insurance has high costs), it obviously does not meet the user's needs, and users are unwilling to purchase such products.

[0003] In order to recommend products that truly meet users' needs, relevant technologies will obtain user behavior information, and then input the user behavior information into a set recommendation algorithm model to obtain business products that meet user needs recommended by the recommendation algorithm model.

[0004] However, recommendation algorithms generally have a T+1 validity period, so there is at least a T+1 delay in the recommendation process. After this T+1 delay, the user's needs may have changed, meaning the recommended product or service they receive may no longer match their actual needs. Summary of the Invention

[0005] The present invention provides a service recommendation method, device, equipment and storage medium based on user behavior, which can solve the technical problem in the prior art that due to the recommendation delay when recommending service products to users, the recommended service products no longer meet the actual needs of users.

[0006] In a first aspect, the present invention provides a service recommendation method based on user behavior, comprising:

[0007] Obtain user behavior information of target users for target business;

[0008] Analyze the user behavior information and generate real-time search data of the user;

[0009] Obtaining user characteristic information of the target user;

[0010] Inputting the real-time search data and the user feature information into a vectorized recall model to obtain a set of candidate services;

[0011] According to a preset user preference parameter, the scores of the target user for the services in the candidate service set are evaluated and sorted to obtain a sorted candidate service sequence set;

[0012] The services ranked before the preset number in the candidate service sequence are used as the optimal recommended services for the target user.

[0013] In a second aspect, the present invention provides a service recommendation device based on user behavior, comprising:

[0014] The first acquisition module is used to obtain user behavior information of the target user for the target business;

[0015] A generation module, configured to analyze the user behavior information and generate real-time search data of the user;

[0016] A second acquisition module is used to obtain user feature information of the target user;

[0017] A service output module, configured to input the real-time search data and the user feature information into a vectorized recall model to obtain a set of candidate services;

[0018] An evaluation module, configured to evaluate the target user's scores for the services in the candidate service set and sort them according to a preset user preference parameter, thereby obtaining a sorted candidate service sequence set;

[0019] The service recommendation module is configured to select the service that is ranked before a preset number in the candidate service sequence as the optimal recommended service for the target user.

[0020] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned user behavior-based service recommendation method when executing the computer program.

[0021] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned service recommendation method based on user behavior are implemented.

[0022] The aforementioned user behavior-based service recommendation method, apparatus, device, and storage medium implement the solution. The service recommendation algorithm model in the related art inputs user behavior information and outputs the optimal recommended service for the target user. This is an end-to-end model that requires a multi-layer neural network model for implementation. The time required to output the optimal recommended service increases exponentially with the feature dimension and number of network layers in the model. In contrast, in this solution, a vectorized recall model is directly utilized to recommend services to users. The vectorized recall model operates by performing vector similarity calculations and an approximate nearest neighbor search algorithm, requiring only simple operations. Therefore, it outputs recommended services faster than an end-to-end model. Furthermore, since the vectorized recall model is less accurate than an end-to-end model, using the vectorized recall model involves roughly screening the set of services that meet user needs. Then, based on user preference parameters, a precise analysis is performed to obtain the optimal recommended service that meets the user's needs. This process of precisely analyzing the optimal recommended service based on user preference parameters is also a simpler analysis process than an end-to-end model, resulting in a faster analysis speed. In summary, the solution of the present invention can quickly and accurately output recommended services that meet user needs, thereby solving the technical problem in the prior art that due to the recommendation delay when recommending service products to users, the recommended service products no longer meet the user's actual needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0024] Figure 1 This is a flow chart of a method for service recommendation steps based on user behavior in one embodiment of the present invention;

[0025] Figure 2 yes Figure 1 A flow chart of step S120;

[0026] Figure 3 yes Figure 2 A flow chart of step S121;

[0027] Figure 4 This is another flowchart of a method for service recommendation based on user behavior in one embodiment of the present invention;

[0028] Figure 5 This is another flowchart of a method for service recommendation based on user behavior in one embodiment of the present invention;

[0029] Figure 6 This is a schematic structural diagram of a service recommendation device based on user behavior in one embodiment of the present invention;

[0030] Figure 7 is a structural diagram of a computer device in one embodiment of the present invention;

[0031] Figure 8 FIG. 2 is another structural diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0033] Figure 1 A flowchart of a method for executing a specified step provided in an embodiment of the present invention, such as Figure 1 As shown, the method for executing a specified step provided by an embodiment of the present invention includes the following steps.

[0034] Step S110: Obtain user behavior information of the target user for the target service.

[0035] Specifically, in this step, the user behavior information is the relevant behavior information of the target user for the target business. For example, the user behavior information can be the user behavior of the user in the relevant target business software (such as collecting, liking, or long-term browsing of a business product in the target business software), or it can be the user's search for a business product in the search box of the relevant target business software. As a specific example, when the target business is property insurance business, the user behavior information can be the user's search for car insurance types, click-through browsing of vehicle accident insurance, and search for car accident insurance premiums in the insurance software; when the target business is medical health insurance business, the user behavior information can be the user's search for the compensation conditions of critical illness insurance or the search for the compensation amount of critical illness insurance in the insurance software.

[0036] Optionally, in this step, in order to obtain user behavior information of the target user, technical tracking points (in the form of plug-ins) can be embedded in business-related platforms (such as business software) in advance. For example, technical tracking points can be embedded in the search box of the business software, or in a specific business product in the business software. The backend server can obtain user behavior information in the technical tracking points after the technical tracking points are triggered by user behavior (that is, the user searches for a preset business product or the user clicks on a business product, etc.).

[0037] Step S120: Analyze the user behavior information to generate real-time search data of the user.

[0038] It should be noted that the user's real-time search data in this step is the data that the user is concerned about. For example, the data that the user searches for in the search bar of the business software is the data that the user is concerned about. The user's click on a business product in the business software indicates that the business product is also the data that the user is concerned about. They can all be considered as the user's real-time search data.

[0039] Specifically, the user's real-time search data may be presented in any form, for example, it may be a search keyword, so that the user can quickly process the real-time search data.

[0040] In some embodiments of the present invention, Figure 2 As shown, step S120 includes the following steps.

[0041] Step S121 : Process the user behavior information according to a stream processing method to generate real-time search data of the user.

[0042] Specifically, stream processing is a technology that collects, processes, and analyzes real-time, unordered, and continuous data sequences (its fundamental principle is to process data as it is generated). For example, Flink utilizes stream processing to process data. Therefore, it's understandable that stream processing of user behavior information can provide real-time user search data, which, to a certain extent, avoids delays in making business recommendations.

[0043] Specifically, in this step, when processing user behavior information, the frequency of search keywords or user clicks can be counted according to the time window. The search history of a single user can also be tracked to count the number of searches, average search interval, and commonly used keywords, so as to obtain the user's real-time search data.

[0044] In some embodiments of the present invention, Figure 3 As shown, step S121 includes the following steps.

[0045] Step S1211: Process the user behavior information according to the stream processing method to generate user search data;

[0046] Step S1212, obtaining the timestamp corresponding to the user behavior information;

[0047] Step S1213: setting a weight corresponding to the user search data according to the time distance between the timestamp corresponding to the user behavior information and the current timestamp.

[0048] The weight is inversely proportional to the time distance corresponding to the timestamp.

[0049] It should be noted that, in step S1212 , there may be multiple pieces of user behavior information. When there are multiple pieces of user behavior information, the timestamp of each piece of user behavior information may be determined.

[0050] Specifically, in step S1213, the weight of the user search data is inversely proportional to the event distance between its corresponding timestamp and the current event stamp, which means that the closer the acquisition time (i.e., timestamp) of the user behavior information corresponding to the user search data is to the current time, the greater its weight, and the farther from the current time, the smaller its weight. That is, the newer the user behavior information, the greater the weight of its corresponding data search data, and the older the user behavior information, the smaller the weight of its corresponding data search data. It can be understood that the newer the user behavior information, the more it can reflect the real needs of the current user, and the older the user behavior information, the more it may deviate from the real needs of the current user. Therefore, when the weight of the newer user search data is set to be larger and the weight of the older user search data is set to be smaller, when the user search data with weights is input into the subsequent vectorized recall model, the business recommended by the vectorized recall model can be more in line with the real-time needs of the user.

[0051] Step S130: Acquire user characteristic information of the target user.

[0052] Specifically, in this step, user characteristic information refers to characteristic information that can reflect the user's consumption habits for the target service. For example, user characteristic information can include information about the user's historically high-frequency service products, as well as the user's historical upper and lower limits on spending. For example, in the insurance sector, such as property insurance, this information can include the longest-term auto insurance purchase by the user, or the auto insurance purchase with the most purchases. In the healthcare insurance sector, this information can include the maximum and minimum amounts of health insurance purchased by the user.

[0053] In some embodiments of the present invention, step S130 includes:

[0054] In a pre-stored user offline feature information database, the user offline feature information corresponding to the identity information of the target user is matched as the user feature information of the target user.

[0055] The user offline feature information in the user offline feature information database is associated with corresponding user identity information.

[0056] Specifically, the identity information of the target user may also be marked in the user search data. Thus, when the user search data is obtained, the user feature information corresponding to the identity information may be found in the user offline feature information database based on the marked identity information.

[0057] It is understandable that since the user feature information is offline, when the user feature information is needed, the processed user feature information can be directly obtained without further processing the user feature information online, thereby speeding up data processing and enabling the final recommended business to be obtained faster.

[0058] Step S140: Input the real-time search data and the user feature information into a vectorized recall model to obtain a set of candidate services.

[0059] Specifically, after the real-time search data and user feature information are input into the vectorized recall model, the vectorized recall model will convert the real-time search data and user feature information into dense vectors (wherein, the real-time search data can be converted into vectors through the word embedding model, and the user feature information can be generated into vectors through feature encoding). The vectors formed by the two can be referred to as user vectors here. The vectorized recall model can calculate the similarity (such as cosine similarity and inner product) between the user vector and the business vector (that is, the business vector corresponding to each business in the full business set) to screen out the alternative business set with the highest similarity to the user vector (that is, the most highly matched with user needs) from the full business set.

[0060] As a specific example, a vectorized recall model can calculate the similarity between user vectors and insurance product vectors, and filter out a set of alternative insurance products that closely match user needs from the entire insurance product set. Products in the alternative insurance set can include auto insurance, health insurance, or accident insurance.

[0061] In some embodiments of the present invention, the vectorized recall model is an Item2Vec model.

[0062] Specifically, the Item2Vec model can transform users' real-time needs and insurance products into association calculations in vector space. It can not only efficiently recall potential businesses, but also capture the semantic associations between businesses and users' dynamic interests, laying the foundation for subsequent refined recommendations.

[0063] Step S150 : evaluating the scores of the target user for the services in the candidate service set according to a preset user preference parameter and sorting the scores to obtain a sorted candidate service sequence set.

[0064] Specifically, in this step, the user preference parameter is a parameter that reflects the user's preference for a service. For example, it can be the historical frequency of recommendations for a service or the user's transaction completion rate for each service. Thus, based on the user preference parameter, the scores of services in the candidate service set are evaluated and sorted, so that services with higher user preference are placed at the top.

[0065] Step S160: Sort the services in the candidate service sequence that are located before the preset number of digits as the optimal recommended services for the target user.

[0066] It is understood that the service recommendation algorithm model in the related art inputs user behavior information and outputs the optimal recommended service for the target user. This is an end-to-end model that requires a multi-layer neural network model to implement. The time it takes to output the optimal recommended service increases exponentially with the feature dimension and number of network layers in the model. In contrast, in this solution, a vectorized recall model is directly utilized to recommend services to users. The vectorized recall model operates by performing vector similarity calculations and an approximate nearest neighbor search algorithm, which requires only simple operations. Therefore, it outputs recommended services faster than end-to-end models. In addition, because the accuracy of the vectorized recall model is not as good as that of the end-to-end model, the vectorized recall model is used to roughly screen the service set that meets the user's needs. Then, based on the user's preference parameters, a precise analysis is performed to obtain the optimal recommended service that meets the user's needs. It is also understood that this process of accurately analyzing the optimal recommended service based on the user's preference parameters is a simpler analysis process than the end-to-end model, and therefore the analysis speed is also faster. In summary, the solution of the present invention can quickly and accurately output recommended services that meet user needs, thereby solving the technical problem in the prior art that due to the recommendation delay when recommending service products to users, the recommended service products no longer meet the user's actual needs.

[0067] It should be noted that, in this step, the number of preset bits can be set according to specific needs of the application, for example, it can be one, four or five.

[0068] In some embodiments of the present invention, Figure 4 As shown, step S160 also includes the following steps.

[0069] Step S171: Push the optimal recommended service to the target user;

[0070] Step S172: Obtain the target user's interactive behavior with respect to the optimal recommended service to form new user behavior data of the target user;

[0071] Step S173: inputting the newly added user behavior data and the user feature information into the vectorized recall model to obtain the updated candidate service set;

[0072] Step S174: Analyze and obtain an updated optimal recommended service for the target user based on the updated set of candidate services.

[0073] Specifically, in step S172, the target user's interactive behavior with respect to the optimal recommended service is the newly added user behavior data on the optimal recommended service. For example, the behavior data may be clicks, browsing, or purchases, etc. It can be understood that these newly added user behavior data can reflect the target user's preference for the optimal recommended service. Therefore, after inputting the newly added behavior data into the quantitative recall model, the obtained updated set of alternative services can further filter the services preferred by the target user. Specifically, in step S174, the step of analyzing and obtaining the updated optimal recommended service for the target user based on the updated set of alternative services can refer to the technical features in steps S150 and S160, and will not be described in detail here.

[0074] In some embodiments of the present invention, Figure 5 As shown, after step S160, the following steps are performed.

[0075] Step S181, analyzing the activity of the target user based on the user behavior information;

[0076] Step S182: when the activity level of the target user is greater than a preset value, determining the target user as a hot user;

[0077] Step S183, when the target user is a hot user, the identity information of the target user, the optimal recommended service of the target user, and the real-time search data are associated to form an associated recommendation strategy corresponding to the target user;

[0078] Step S184: storing the associated recommendation strategy, the identity information of the target user, and the mapping relationship between the associated recommendation strategy and the identity information of the target user in a preset memory.

[0079] Specifically, for step S181, when analyzing the activity of the target user based on user behavior information, it can be determined based on user behavior information such as the target user's login time or number of clicks on the business platform of the relevant target business. It can be understood that the longer the target user's login time on the business platform, the more interested they are in the target business. The more times the target user clicks on the business in the business platform, the more interested they are in the target business, that is, their activity is greater, and it is easier for them to purchase business products.

[0080] For this type of user, since they are more likely to purchase business products, the optimal recommended business, their identity information and real-time search data can be formed into an associated recommendation strategy and stored in the memory. In this way, when the user wants to purchase a specific business product, the recommended business product can be quickly obtained from the memory based on their identity information and real-time search data, without the need to re-analyze it online based on their real-time search data and feature information, which can effectively save the user's time and improve the user experience.

[0081] It should be understood that the order of execution of the steps in the above embodiments does not necessarily imply a specific order of execution. The order of execution of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention. The software tools or components not provided by our company that appear in the embodiments of this application are merely examples and do not represent actual use.

[0082] In one embodiment, a service recommendation device based on user behavior is provided, which corresponds to the service recommendation method based on user behavior in the above embodiment. Figure 6 As shown, the analysis device includes a first acquisition module 610, a generation module 620, a second acquisition module 630, a business output module 640, an evaluation module 650 and a business recommendation module 660. The functional modules are described in detail as follows:

[0083] The first acquisition module 610 is used to obtain user behavior information of the target user for the target service;

[0084] A generating module 620 is configured to analyze the user behavior information and generate real-time search data of the user;

[0085] The second acquisition module 630 is used to obtain user feature information of the target user;

[0086] A service output module 640 is configured to input the real-time search data and the user feature information into a vectorized recall model to obtain a set of candidate services;

[0087] An evaluation module 650 is configured to evaluate the target user's scores for the services in the candidate service set and sort them according to a preset user preference parameter to obtain a sorted candidate service sequence set;

[0088] The service recommendation module 660 is configured to select the service that is ranked before a preset number in the candidate service sequence as the optimal recommended service for the target user.

[0089] In one embodiment, the second acquisition module 630 is specifically configured to:

[0090] In a pre-stored user offline feature information database, user offline feature information corresponding to the identity information of the target user is matched as user feature information of the target user, wherein the user offline feature information in the user offline feature information database is associated with corresponding user identity information.

[0091] In one embodiment, the business recommendation module 660 is further configured to:

[0092] Pushing the optimal recommended service to the target user;

[0093] Obtaining the target user's interactive behavior with respect to the optimal recommended service to form new user behavior data of the target user;

[0094] Inputting the newly added user behavior data and the user feature information into the vectorized recall model to obtain the updated candidate service set;

[0095] An updated optimal recommended service for the target user is obtained by analysis based on the updated set of candidate services.

[0096] In one embodiment, the vectorized recall model is an Item2Vec model.

[0097] In one embodiment, the generating module 620 is specifically configured to:

[0098] The user behavior information is processed according to a stream processing method to generate real-time search data of the user.

[0099] In one embodiment, the generating module 620 is further configured to:

[0100] Process the user behavior information according to the stream processing method to generate user search data;

[0101] Obtaining a timestamp corresponding to the user behavior information;

[0102] The weight corresponding to the user search data is set according to the time distance between the timestamp corresponding to the user behavior information and the current timestamp, wherein the weight is inversely proportional to the time distance corresponding to the timestamp.

[0103] In one embodiment, the business recommendation module 660 is further configured to:

[0104] Analyzing the activity of the target user based on the user behavior information;

[0105] When the activity level of the target user is greater than a preset value, determining the target user as a hot user;

[0106] When the target user is a hot user, the identity information of the target user, the optimal recommended service of the target user and the real-time search data are associated to form an associated recommendation strategy corresponding to the target user;

[0107] The associated recommendation strategy, the identity information of the target user, and a mapping relationship between the associated recommendation strategy and the identity information of the target user are stored in a preset memory.

[0108] The present invention provides a service recommendation device based on user behavior. The service recommendation algorithm model in the related art inputs user behavior information and outputs the optimal recommended service for the target user. This is an end-to-end model that requires a multi-layer neural network model for implementation. The time it takes to output the optimal recommended service increases exponentially with the feature dimensions and number of network layers in the model. In this solution, a vectorized recall model is directly utilized to recommend services to users. The vectorized recall model performs vector similarity calculations and an approximate nearest neighbor search algorithm, requiring only simple operations. Therefore, it outputs recommended services faster than an end-to-end model. Furthermore, because the vectorized recall model is less accurate than an end-to-end model, the vectorized recall model is used to roughly screen a set of services that meet user needs. Then, based on user preference parameters, an accurate analysis is performed to obtain the optimal recommended service that meets the user's needs. This process of accurately analyzing the optimal recommended service based on user preference parameters is also a simpler analysis process than an end-to-end model, resulting in a faster analysis speed. In summary, the solution of the present invention can quickly and accurately output recommended services that meet user needs, thereby solving the technical problem in the prior art that due to the recommendation delay when recommending service products to users, the recommended service products no longer meet the user's actual needs.

[0109] The specific definition of the user behavior-based service recommendation device can be found in the definition of the user behavior-based service recommendation method above and will not be repeated here. Each module in the above-mentioned user behavior-based service recommendation device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor of the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each of the above modules.

[0110] Based on the above business recommendation method based on user behavior, such as Figure 7 As shown, an embodiment of the present invention further provides a schematic structural diagram of a device for performing steps of service recommendation based on user behavior, the device comprising a processor 71 and a memory 72 coupled to the processor 71. The memory 72 stores a computer program, which, when executed by the processor 71, causes the processor 71 to perform the steps of the service recommendation method based on user behavior in the above embodiment.

[0111] For other details about how the processor 71 in the device for the above-mentioned user behavior-based service recommendation step implements the above-mentioned technical solution, please refer to the description of the user behavior-based service recommendation method provided in the above-mentioned invention embodiment, which will not be repeated here.

[0112] Among them, the processor 71 can also be called a CPU (Central Processing Unit), and the processor 71 may be an integrated circuit chip with signal processing capabilities; the processor 71 can also be a general-purpose processor, DSP (Digital Signal Process), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, among which the general-purpose processor can be a microprocessor or the processor 71 can also be any conventional processor, etc.

[0113] like Figure 8As shown, an embodiment of the present invention further provides a schematic diagram of the structure of a computer-readable storage medium, on which a readable computer program 81 is stored; wherein, the computer program 81 can be stored in the above-mentioned storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk, a ROM (Read-Only Memory), a RAM (Random Access Memory), and other media that can store program code, or a terminal device such as a computer, server, mobile phone, or tablet.

[0114] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or module, which can be electrical, mechanical or other forms.

[0115] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of the present embodiment according to actual needs.

[0116] In addition, the functional modules in the various embodiments of the present invention may be integrated into a single processing module, each module may exist physically separately, or two or more modules may be integrated into a single module. The integrated modules may be implemented in the form of hardware or software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may be stored in a computer-readable storage medium.

[0117] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware, or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0118] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium), or a semiconductor medium (e.g., an SSD (solid state disk)).

[0119] The technical solution provided by the present invention is introduced in detail above. Specific examples are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

[0120] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, optical storage, etc.) containing computer-usable program code.

[0121] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0122] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0123] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0124] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A service recommendation method based on user behavior, characterized in that: include: Obtain user behavior information of target users for target business; Analyze the user behavior information and generate real-time search data of the user; Obtaining user characteristic information of the target user; Inputting the real-time search data and the user feature information into a vectorized recall model to obtain a set of candidate services; According to a preset user preference parameter, the scores of the target user for the services in the candidate service set are evaluated and sorted to obtain a sorted candidate service sequence set; The services ranked before the preset number in the candidate service sequence are used as the optimal recommended services for the target user.

2. The service recommendation method based on user behavior according to claim 1, characterized in that: The acquiring of user characteristic information of the target user comprises: In a pre-stored user offline feature information database, user offline feature information corresponding to the identity information of the target user is matched as user feature information of the target user, wherein the user offline feature information in the user offline feature information database is associated with corresponding user identity information.

3. The service recommendation method based on user behavior according to claim 1, characterized in that: After sorting the services that are ranked before the preset number in the candidate service sequence as the optimal recommended services for the target user, the method further includes: Pushing the optimal recommended service to the target user; Obtaining the target user's interactive behavior with respect to the optimal recommended service to form new user behavior data of the target user; Inputting the newly added user behavior data and the user feature information into the vectorized recall model to obtain the updated candidate service set; An updated optimal recommended service for the target user is obtained by analysis based on the updated set of candidate services.

4. The service recommendation method based on user behavior according to claim 1, characterized in that: The vectorized recall model is the Item2Vec model.

5. The service recommendation method based on user behavior according to claim 1, characterized in that: Analyzing the user behavior information to generate real-time search data of the user includes: The user behavior information is processed according to a stream processing method to generate real-time search data of the user.

6. The service recommendation method based on user behavior according to claim 5, characterized in that: The step of processing the user behavior information in a stream processing manner to generate the user's real-time search data includes: Process the user behavior information according to the stream processing method to generate user search data; Obtaining a timestamp corresponding to the user behavior information; The weight corresponding to the user search data is set according to the time distance between the timestamp corresponding to the user behavior information and the current timestamp, wherein the weight is inversely proportional to the time distance corresponding to the timestamp.

7. The service recommendation method based on user behavior according to claim 1, characterized in that: After sorting the services that are ranked before the preset number in the candidate service sequence as the optimal recommended services for the target user, the method further includes: Analyzing the activity of the target user based on the user behavior information; When the activity level of the target user is greater than a preset value, determining the target user as a hot user; When the target user is a hot user, the identity information of the target user, the optimal recommended service of the target user and the real-time search data are associated to form an associated recommendation strategy corresponding to the target user; The associated recommendation strategy, the identity information of the target user, and a mapping relationship between the associated recommendation strategy and the identity information of the target user are stored in a preset memory.

8. A service recommendation device based on user behavior, characterized in that: include: The first acquisition module is used to obtain user behavior information of the target user for the target business; A generation module, configured to analyze the user behavior information and generate real-time search data of the user; A second acquisition module is used to obtain user feature information of the target user; A service output module, configured to input the real-time search data and the user feature information into a vectorized recall model to obtain a set of candidate services; An evaluation module, configured to evaluate the target user's scores for the services in the candidate service set and sort them according to a preset user preference parameter, thereby obtaining a sorted candidate service sequence set; The service recommendation module is configured to select the service that is ranked before a preset number in the candidate service sequence as the optimal recommended service for the target user.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the service recommendation method based on user behavior according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the service recommendation method based on user behavior according to any one of claims 1 to 7 are implemented.

Images