Business data processing, information recommendation method and device, electronic equipment and medium

Abstract

The application provides a service data processing method and device, information recommendation method and device, electronic equipment and medium. According to the embodiment of the application, efficient and accurate analysis of user demand is realized. The service data processing method comprises the following steps: determining service type information used by a user and use behavior information of the user according to record data of service related behaviors of the user; obtaining related behavior change information of the user on a preset service type within a preset period according to the service type information and the use behavior information; and obtaining demand degree information of the user on the service of the preset service type according to the related behavior change information of the user on the preset service type within the preset period.

Description

Technical Field

[0001] This application relates to the field of information recommendation technology, and in particular to a business data processing, information recommendation method, apparatus, electronic device and medium. Background Technology

[0002] With the development of computer technology, people's lives, including clothing, food, housing, and transportation, have become significantly intertwined with it. Especially with the advancement of mobile terminal technology and the increasing diversity of applications, user needs can be met through various means connected to mobile applications or other computer products. To improve the user experience with computer products, it is necessary to analyze and predict user demands for various types of services. However, in general, systematic analysis and prediction of user needs requires significant time and hardware / software costs, and the accuracy is relatively low. Summary of the Invention

[0003] This application provides a business data processing, information recommendation method, apparatus, electronic device, and medium to achieve more accurate user demand prediction information at a lower cost.

[0004] In a first aspect, embodiments of this application provide a business data processing method, comprising: determining, based on recorded data of a user's business-related behaviors, information on the type of business used by the user and information on the user's usage behavior; obtaining information on changes in the user's related behaviors toward a preset business type within a preset period based on the business type information and the usage behavior information; and obtaining information on the degree of user demand for the preset business type based on the information on changes in the user's related behaviors toward the preset business type within a preset period.

[0005] Secondly, embodiments of this application provide an information recommendation method, comprising: determining target business information required by the user based on the user's demand level information for a preset business type; wherein the demand level information is generated by the business data processing method provided in any embodiment of this application; and generating recommendation information based on the target business information.

[0006] Thirdly, embodiments of this application provide a business data processing apparatus, comprising: a data recording and analysis module, configured to determine, based on recorded data of user business-related behaviors, information on the type of business used by the user and information on the user's usage behavior; a related behavior change information module, configured to obtain information on changes in the user's related behavior towards a preset business type within a preset period, based on the business type information and the usage behavior information; and a demand level information module, configured to determine information on the user's demand level for the preset business type based on the information on changes in the user's related behavior towards the preset business type within a preset period.

[0007] Fourthly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor implements any of the methods described above when executing the computer program.

[0008] Fifthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the methods described above.

[0009] Compared with the prior art, this application has the following advantages:

[0010] It can determine the duration for which a target texture has not been invoked based on the latest invocation time of the target texture. If the duration for which a target texture has not been invoked exceeds a set duration threshold, the target texture will be deleted from the display storage space, thereby saving display storage space resources and avoiding problems such as lag and forced closure of map applications caused by excessive display storage space occupation.

[0011] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application, it can be implemented according to the contents of the specification. In order to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0012] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments according to this application and should not be construed as limiting the scope of this application.

[0013] Figure 1 This is a schematic diagram illustrating an application scenario of the business data processing method according to an embodiment of this application;

[0014] Figure 2 This is a schematic diagram of the business data processing method according to an embodiment of this application;

[0015] Figure 3 This is a schematic diagram illustrating a specific example of a business data processing method in this application;

[0016] Figures 4A-4F A schematic diagram of the centroid vector as an example of this application;

[0017] Figure 5 This is a schematic diagram illustrating feature analysis of one example of this application;

[0018] Figure 6 This is a schematic diagram of user requirement data according to one embodiment of this application;

[0019] Figure 7 This is a schematic diagram of the interface of an information recommendation method according to an embodiment of this application;

[0020] Figure 8 This is a schematic diagram of a business data processing apparatus according to an embodiment of this application; and

[0021] Figure 9 This is a block diagram of an electronic device used to implement embodiments of this application. Detailed Implementation

[0022] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the concept or scope of this application. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0023] To facilitate understanding of the technical solutions of the embodiments of this application, the relevant technologies of the embodiments of this application are described below. The following relevant technologies are optional solutions and can be combined with the technical solutions of the embodiments of this application in any way, and all of them fall within the protection scope of the embodiments of this application.

[0024] Figure 1 This is a schematic diagram illustrating an application scenario for implementing the method of the embodiments of this application. For example... Figure 1 As shown, the business data processing method of this application embodiment can be applied to a system including a user terminal and a server 101. The user terminal may include at least one, or multiple user terminals belonging to the same user. For example, as... Figure 1 As shown, multiple user terminals may include mobile phones 102, laptops 103, tablets 104, and desktop computers 105. These user terminals can be identified as belonging to the same user through certain methods. For example, the same user logs into these user terminal devices using the same user account and uses the computer products to perform business data-related operations. Specifically, business data-related operations may include using a webpage to query business-related information, using a webpage to reserve business services, using a webpage to pay for or prepay for reserved services, using a computer application to query business-related information, using an application to reserve business services, using an application to pay for or prepay for reserved services, and publishing evaluation information related to business information, etc.

[0025] When a user performs business-related operations through a user terminal, record data of the business-related behavior is generated, such as record data of purchasing vouchers or booking hotels. The terminal can send the record data to the server 101, and the server 101 can also actively retrieve the record data from the terminal. After obtaining the record data of business-related behavior, the server 101 can process the record data using the business data processing method provided in any embodiment of this application to determine the user's degree of demand for the preset business type.

[0026] The server 101 can further generate an information recommendation strategy based on the user's demand for services of preset service types. Based on the information recommendation strategy and the recommendable data in the database, it generates recommended information and sends it to the user terminal, so that the user can receive the recommended information through the user terminal, which helps the user select the recommended information related to the service they need and shortens the time the user spends filtering the information they need.

[0027] The services provided in this application embodiment can include various preset service types such as catering, shopping, medical care, travel services, tourism, hotels, and entertainment. The service data in this application embodiment can be data generated by users performing operations related to preset service types using their user terminals. All user-related data used in all embodiments of this application were obtained with the user's permission or direct authorization, and the data acquisition process does not infringe upon the user's right to information and complies with relevant laws and regulations.

[0028] The business data processing method provided in the embodiments of this application, such as Figure 2 As shown, it can be used to process map data during the operation of map applications, including steps S201-S203.

[0029] In step S201, the user's service type information and user behavior information are determined based on the recorded data of the user's business-related behavior.

[0030] In this embodiment, the recorded data of a user's business-related behavior can be network or computer record data generated when a user uses a computer product to perform business-related operations. For example, a user using an electronic map to search for a restaurant generates map search record data. Similarly, a user using a ride-hailing app to hail a ride to a restaurant generates ride-hailing record data. Furthermore, a user using a payment app to complete a payment at a restaurant generates payment records. Finally, a user using a review app to search for reviews of a restaurant generates review records.

[0031] In this embodiment, the recorded data of a user's business-related behavior may include the recorded data generated when a user uses a computer product to perform operations related to one of a plurality of preset business types. For example, the preset business types may include catering, hotels, and entertainment. Therefore, when a user queries data related to any one of the catering, hotel, or entertainment businesses, the resulting recorded data can be used as the recorded data of the user's business-related behavior.

[0032] In this embodiment, the service type information used by the user can include the degree of association with each preset service type. Since user-generated service-related behaviors using computer products may involve more than one preset service type, for example, if a user checks into a hotel and dines at a tourist attraction, the resulting record data may be related to three preset service types: tourism, hotel, and catering. Therefore, for each different type of user record data, the service type information used by the user can be determined individually.

[0033] For example, if a user makes a hotel reservation using a webpage, application, or other computer-related product, the resulting record data can be determined to have a 100% correlation with the preset business type "hotel," and a 0% correlation with other preset business types. Accordingly, the business type information could include: Hotel 100%, Other 0%.

[0034] For example, if a user uses computer-related products such as web pages or applications to check into a hotel within a scenic area, the resulting record data can be determined to have a 90% correlation with the preset business type "tourism," a 10% correlation with the business type "hotel," and a 0% correlation with other preset business types. Accordingly, the business type information could include: Tourism 90%, Hotel 10%, Other 0%.

[0035] In this application embodiment, when the recorded information of user behavior related to business involves two or more preset business types, the primary and secondary purposes of the user's behavior can be determined based on the recorded information of user behavior related to business, thereby determining the degree of relevance to different preset business types, which is used as business type information.

[0036] In this embodiment, the usage behavior information may include at least one of the following: behavior method, behavior type, or behavior attribute. For example, if a user uses a webpage, application, or other computer-related product to search for a hotel, and it is determined that the hotel searched by the user is located within a scenic area, then the usage behavior information can be determined to include: method - online, type - search, and attribute - tourist accommodation.

[0037] In another embodiment of this application, the behavioral information may include the user behavior reflected in the recorded information and the type to which it belongs among a preset plurality of behavioral types. For example, the preset plurality of behavioral types may include: online transaction type, online non-transaction type, offline transaction type, or offline non-transaction type, etc.

[0038] In another embodiment of this application, the use of behavioral information may further include the type of user behavior among a preset plurality of behavioral types determined based on the recorded information, and further calculation and analysis of other information related to user behavior. For example, the type of user behavior among the preset plurality of behavioral types is online transaction type, and the corresponding type number is 001.

[0039] In step S202, based on the service type information and usage behavior information, information on changes in the user's behavior towards the preset service type within a preset period is obtained.

[0040] In this embodiment, the preset period can be determined based on the update cycle of different businesses. For example, the update cycle for tourism-related behaviors is one month, while catering behaviors occur every day in users' lives, so the update cycle for catering-related behaviors is one day.

[0041] In another embodiment of this application, the preset period can be a single period, such as one day.

[0042] In another embodiment of this application, the preset period can be multiple periods, such as one day, one week, and one month. When there are multiple preset periods, the user's behavioral change information related to the preset service type within the preset period can include behavioral change information within at least one preset period.

[0043] In this embodiment of the application, obtaining relevant behavioral change information of a user on a preset business type within a preset period based on business type information and usage behavior information may include determining the behavioral change period and relevant business type of the user within multiple preset periods based on business type information and usage behavior information, and then further determining the behavioral change information of the relevant business type within the behavioral change period.

[0044] For example, preset periods can include: one day, one week, and one month. Preset business types can include: catering, hotels, and tourism. Based on business type information and usage behavior information, the changes in users' tourism-related behaviors within a week are determined to be stable.

[0045] In step S203, based on the user's behavioral change information related to the preset service type within a preset period, information on the user's demand for the preset service type is obtained.

[0046] In this embodiment, the user's demand information for services of preset service types can include the user's demand information for each of multiple preset service types. The demand information can be represented by a numerical score; for example, a score between 0 and 1 can be used to represent the user's demand information for a single preset service type. Therefore, the user's demand information for services of preset service types can include: a demand score of 0.1 for the first preset service type; a demand score of 0.9 for the second preset service type; and a demand score of 0.8 for the third preset service type.

[0047] In this embodiment, the system can analyze the user's demand for a specific type of service within a set period based on the recorded data of the user's business-related behaviors. This allows the system to collect data on the user's network usage without investing a large amount of additional hardware and software resources or conducting special user surveys, thereby reducing the cost of user demand analysis and improving its efficiency.

[0048] In one embodiment of this application, the business data processing method further includes: when a user uses an application to perform business activities related to points of interest (POIs) in a map application, obtaining recorded data of the user's use of the map application as recorded data of the user's business-related activities.

[0049] In this embodiment, points of interest can be location points in a map application that are relevant to the business. For example, restaurants, shopping malls, tourist attractions, stadiums, bus stops, subway stations, or hotels.

[0050] In this embodiment of the application, user record data can be obtained by the user performing business activities related to points of interest in the map application using map applications or non-map applications, thereby helping to determine the business type to which the record data belongs based on the points of interest.

[0051] The aforementioned non-map applications can be location-related applications, including review apps, food delivery apps, ride-hailing apps, travel service apps, or restaurant service apps specifically designed for reviewing points of interest. While these applications are not map applications, they utilize user location information during operation and may collect user travel trajectories, thus obtaining user data. For example, if a user uses a ride-hailing app to travel from a first point of interest to a second point of interest, and then uses a restaurant service app or review app to purchase a voucher for a restaurant at the second point of interest, the resulting ride-hailing and restaurant voucher purchase records can be considered user data, identifying the business type as either restaurant service or ride-hailing.

[0052] In one embodiment of this application, determining the service type information used by a user based on the recorded data of the user's business-related behavior includes: obtaining the degree of user's use of services of at least one preset service type based on the recorded data of the user's business-related behavior; and using the degree of use information as service type information.

[0053] In this embodiment, the user's usage information for at least one preset service type can include the user's usage information for each of the at least one preset service type. For example, among the seven preset service types, the usage level of the first preset service type is X1, the usage level of the second preset service type is X2, the usage level of the third preset service type is X3, and so on, with the usage level of the seventh preset service type being X7.

[0054] In another implementation, the preset business type can include at least two types in a combined manner. For example, the preset business type can include a combination of at least two types. Specifically, the preset business type could be a scenic area hotel business type, which includes both tourism and hotel types.

[0055] In another implementation, obtaining user usage information for at least one preset service type based on recorded user behavior data can include: obtaining user usage information for at least one preset service type, and preset period information corresponding to the at least one preset service type, based on recorded user behavior data. Using the usage information as service type information can also include using both the usage information and the preset period information as service type information.

[0056] The preset cycle information corresponding to the preset service types can be the same or different. For example, there are 7 preset service types, each with one preset cycle information; or at least one preset service type corresponds to one preset cycle information; or all preset service types correspond to the same preset cycle information.

[0057] In this embodiment, the preset period information may include a preset period corresponding to a service of at least one preset service type.

[0058] In another implementation, obtaining user usage information for at least one preset business type based on recorded user business-related behavior data may include: obtaining user usage information for a merged business type based on recorded user business-related behavior data; determining at least two merged preset business types based on the merged business type; and obtaining user usage information for at least two merged preset business types based on the at least two merged preset business types.

[0059] In real life, there are situations where a single behavior encompasses multiple pre-defined business types (i.e., integrated business types). For example, when a user dines while traveling, their business-related behavior involves both tourism and dining, two pre-defined business types. The above embodiment can break down the integrated business type into more granular pre-defined business types, enabling accurate analysis of situations where a single behavior is related to multiple pre-defined business types.

[0060] In another implementation, obtaining information on the user's usage of at least one preset business type based on the recorded data of the user's business-related behavior may include obtaining information on the user's usage of at least one non-integrated preset business type based on the recorded data of the user's business-related behavior.

[0061] In this embodiment, the usage information of at least one preset business type is used as business type information, thereby enabling the acquisition of the business type associated with user behavior in the recorded data, and the analysis of user behavior through the associated business type.

[0062] In one embodiment of this application, information on the degree of user usage of multiple preset types of services is obtained based on the recorded data of the user's business-related behavior, including: obtaining information on the degree of association between the services used by the user and at least one preset service type; and obtaining the degree of usage information based on the recorded data and the degree of association.

[0063] In this embodiment, the density can be represented by a score; the higher the score, the higher the density.

[0064] In this embodiment, usage level information can be used to comprehensively represent a user's usage type, frequency, and duration of multiple preset service types. Usage type can represent the type of operation a user performs on a service, such as querying service-related information, following service-related information, saving service-related information, and consuming service-related products. For each operation type, weights and other information can be assigned. Based on the operation time corresponding to each different operation type, combined with the weight information, the usage level information is obtained. Usage level information can also be represented by scores or symbols. For example, if a user browses service-related information once for one hour, the usage level information can be 2. Similarly, if a user pays for service-related products twice for a total of three minutes, the usage level information can be 5.

[0065] In this embodiment, usage information is obtained based on recorded data and scores, thereby enabling the determination of service type information based on the usage information.

[0066] In one implementation, determining user behavior information based on recorded data of user business-related behaviors includes: obtaining user behavior type information and user definition information corresponding to the behavior type information based on recorded data of user business-related behaviors; and using the behavior type information and user definition information as user behavior information.

[0067] In this embodiment, the behavior type information can be a pre-defined type. At least one behavior type information can be pre-determined based on the implementation method of the user's participation in or use of the service. Based on the specific content of the recorded information, the behavior type information corresponding to the user's participation in or use of the service in the pre-set at least one behavior type information can be determined.

[0068] In this embodiment of the application, user definition information that corresponds to behavior type information can be set. That is, the way users participate in business can be defined according to the pattern of business participation behavior that users frequently perform. For example, user definition information may include long-term stable users, short-term rapidly increasing users, users during the development period, users during the decline period, users with fluctuating demand, etc.

[0069] In one implementation, user behavior type information is obtained based on recorded data of user business-related behaviors, including at least one of the following: if the recorded data indicates that the user completed a business activity at a physical address, the behavior type information is determined to be an offline type; if the recorded data indicates that the user inquired about business online but did not complete the business activity online, the behavior type information is determined to be an online non-business type; if the recorded data indicates that the user completed a business activity online, the behavior type information is determined to be an online business type.

[0070] In this embodiment, whether a user has completed a business transaction online can be distinguished by whether they have completed business confirmation operations such as payment or reservation. If a user completes the confirmation operation while performing a business-related action online, the recorded data can be confirmed to indicate that the user has completed the business activity online. If a user does not complete the confirmation operation while performing a business-related action online, the recorded data can be confirmed to indicate that the user has not completed the business activity online.

[0071] In this embodiment, information about a user completing business activities at a physical address can be obtained by locating the user's terminal. When determining the user's demand for a preset type of service based on the user's behavior type information, different weights can be assigned to different behavior types. For example, online user behavior is easier to capture and record; therefore, higher weights can be assigned to online non-business types and online business types.

[0072] In this embodiment, a comprehensive analysis of users' online and offline behaviors can be performed, improving the comprehensiveness and accuracy of user demand information results.

[0073] In one implementation, based on business type information and usage behavior information, information on changes in user behavior related to a preset business type within a preset period is obtained, including: analyzing the stability of user usage behavior of the preset business type within a sampling period based on business type information and usage behavior information, and generating time-series features; calculating the distance information between the time-series features and centroid features; the centroid features are a first centroid feature corresponding to stable business behavior within the preset period, a second centroid feature corresponding to upward trend business behavior within the preset period, and a third centroid feature corresponding to downward trend business behavior within the preset period; and determining the information on changes in user behavior related to the preset business type within the preset period based on the distance information.

[0074] In this embodiment, the recorded data within the sampling period can be used to generate time-series features according to the temporal order of unit time within the sampling period. The dimensions of the time-series features can be consistent with the sampling period. For example, recorded data within a 50-day sampling period can generate 50-dimensional time-series features; recorded data within a 30-day sampling period can generate 30-dimensional time-series features; and recorded data within a 40-day sampling period can generate 40-dimensional time-series features.

[0075] In this embodiment, the behavioral change information related to the preset business type may include the trend information of the frequency of occurrence of the relevant behavior of the preset business type and the trend information of the conversion probability of the relevant behavior of the preset business type.

[0076] In this embodiment of the application, the first centroid feature, the second centroid feature, and the third centroid feature are preset reference vectors.

[0077] When there are multiple preset periods, each preset period can be configured with a first centroid feature, a second centroid feature, and a third centroid feature. The first centroid feature is a vector representation of stable business behaviors occurring within the preset period. If the features calculated based on the recorded data within the preset period are consistent with the first centroid feature, it can be determined that the stability of the user's business-related behaviors within the preset period is the highest.

[0078] In one implementation, determining user behavior changes related to a preset service type within a preset period based on distance information includes: determining the stability of user behavior related to the preset service type within a preset period, and using the stability as user behavior change information, when the time-series feature is closest to the first centroid feature among the first, second, and third centroid features; determining the rise in user behavior related to the preset service type within a preset period, and using the rise in user behavior change information, when the time-series feature is closest to the second centroid feature among the first, second, and third centroid features; and determining the fall in user behavior related to the preset service type within a preset period, and using the fall in user behavior change information, when the time-series feature is closest to the third centroid feature among the first, second, and third centroid features.

[0079] In this embodiment, the stability score can be a score representing the degree of stability when the overall behavioral trend is stable. The rise score can be a score representing the degree of rise when the overall behavioral trend is rising. The fall score can be a score representing the degree of fall when the overall behavioral trend is falling.

[0080] In this embodiment, the proximity between temporal features and centroid features is determined by distance information. The calculation process is simple, consumes few computational resources, and has high computational efficiency.

[0081] In one implementation, the sampling period is longer than a preset period.

[0082] In this embodiment, when there are multiple preset periods, the sampling period can be greater than the maximum value among the multiple preset periods. A sampling period greater than the preset period allows for sufficient recorded data to be obtained within the sampling period, resulting in more accurate user demand analysis results.

[0083] In one implementation, the preset period is set to correspond to a preset service type. In this embodiment, the preset period corresponds to the preset service type, so different preset service types can correspond to different preset periods. This allows for setting a preset period that is generally adapted to the frequency of related behaviors of the service type, facilitating a more accurate analysis of user demand information.

[0084] In one implementation, the preset period includes multiple different periods. For example, the preset period includes days, weeks, and months.

[0085] This application also provides a method for training a business data processing model, comprising: inputting training samples into a business data processing model to be trained, causing the business data processing model to be trained to execute the business data processing method provided in any one of the claims of this application to obtain an estimated value of demand level information; and training the business data processing model to be trained based on the reference value of demand level information of the training samples and the estimated value of demand level information to obtain a trained business data processing model.

[0086] In this embodiment, the business data processing model to be trained may include a feature generator, a feature extractor, and a general ranking sub-model. The feature generator transforms input data into feature vectors; the feature extractor compares the feature vectors with preset centroid vectors to obtain a comparison result; and the general ranking sub-model can obtain demand level information based on the comparison result. This embodiment can obtain user demand level information for different types of business by training a single model, reducing the number of models and lowering model deployment and operating costs.

[0087] This application also provides an information recommendation method, comprising: determining target business information needed by the user based on the user's demand level information for a preset business type; the demand level information is generated by any of the business data processing methods provided in this application; and generating recommendation information based on the target business information needed by the user. The aforementioned preset business type may be the preset business type in the foregoing embodiments.

[0088] In this embodiment, the recommendation information may include information that prompts or suggests user behavior. For example, when the user's demand level information indicates a high demand for public transportation, the system can recommend the locations of public transportation facilities such as bus stops and subway stations near the user's location when the user is out and about. Similarly, if a user frequently dine at a certain type of restaurant, the system can recommend the locations of similar restaurants near the user's location when the user is traveling.

[0089] In this embodiment, the recommendation information generation method can be executed on the server side, or jointly by the server and the terminal. For example, recommendation information can be generated on the server side. Alternatively, the server can obtain the user's requested services, and the terminal can obtain these services from the server and generate recommendation information on the terminal.

[0090] This application also provides an information recommendation method, including: obtaining recommendation information based on an event that detects the occurrence of recommendation conditions; the recommendation information is generated by the recommendation information generation method provided in any embodiment of this application; and sending the recommendation information to the user.

[0091] In this embodiment, the event that triggers the recommendation criteria can be an event that satisfies the criteria, such as a user opening the target app, a user leaving their city of residence and arriving in another city, or a user approaching a recommended point of interest within a set range. The recommendation criteria can be determined based on the user's level of need information, or by determining the user's desired business based on that information, and then determining the recommendation criteria accordingly. For example, if a user has a high demand for a certain type of restaurant, and recorded data indicates that the user typically dine at that type of restaurant on weekdays, then recommendation information about that type of restaurant can be sent to the user when their mealtime approaches on weekdays.

[0092] In another embodiment, the recommendation criteria can also be related to the status of the recommended business or to user settings. For example, based on user demand information, if it is determined that a user has a high demand for a certain type of restaurant on weekdays, then recommendation information can be sent to the user when that type of restaurant launches new products. Furthermore, if a user has set their preferences to not recommend or reduce recommendations for a certain type of information, but to recommend another type of information, then information that the user has set as recommendable can be recommended.

[0093] Generally, analyzing or predicting user needs requires collecting a large amount of user information. This approach can be costly and ineffective. The information collected for this purpose consists of objective facts about the user, such as age, gender, family information, and company information. When vertical businesses use this user information, they need to establish a connection between the vertical business and the user information, sometimes requiring manual intervention. Furthermore, because the mapping between user information and user needs across different verticals is indirect, implementation is often ineffective in the context of refined operations. The mapping method may differ from the actual user needs, making it difficult to achieve significant returns. For example, when identifying potential customers for family-friendly hotels, it's generally necessary to select users with children who also have sufficient assets. However, not everyone with children and sufficient assets needs to stay in family-friendly hotels. Therefore, in the current context of refined operations, this approach fails to meet the dual requirements of high coverage and low cost for information recommendation services. Additionally, collecting large amounts of user information can lead to high resource consumption. The collection of user information requires collecting information from multiple dimensions. Since the information from multiple dimensions is not very correlated, it is necessary to design mining or collection schemes for each dimension. Dozens or even hundreds of models need to be built to execute the mining process, which consumes a lot of research, computing and storage resources. In addition, since a large number of models are used in the process of collecting a large amount of user information, the cost of manually maintaining the models is very high.

[0094] The business data analysis method provided in this application proposes a general, all-category user demand analysis model based on user behavior time-series. This approach has low business application costs, transforming the collected information from user information into records of user business-related behaviors. This provides a more direct link between vertical businesses and the collected records, reducing the time consumed in the business-user information association stage and improving the implementation effectiveness of various types of businesses. Furthermore, the business data analysis method provided in this application has low resource consumption. This application can use a single general model, single-model system, or module to produce user information analysis results for all categories of industries, reducing data storage and computational complexity, while also reducing manual maintenance costs. Moreover, this general model or module remains applicable to cold-start business scenarios that do not yet have user feedback data.

[0095] In one example of this application, a user demand analysis model can be constructed. User behavior time-series data, i.e., the recorded data in the aforementioned embodiments, is input into the analysis model. The analysis model outputs the degree of user demand for all categories of life service-related businesses based on the recorded data. The output data of the analysis model is refined to the degree of business demand. For a specific life service business, the analysis model further stratifies the demand for the user-selected business type (i.e., the preset business type and / or preset business type in the aforementioned embodiments) based on the relationship between the user and the platform. In this example, the input-output expression of the analysis model is: Analysis Model (User Data, Business Scope, Business Level, Demand Cycle) = Demand Degree.

[0096] In this example: the output data of the analysis model is the degree of user demand for a preset type of service, which can be expressed as SOD (Score of Demand), representing the degree of match between the user and a certain demand of a certain service type, which can be equivalent to the degree of demand information in the previous embodiment.

[0097] In this example, the input data of the analysis model is equivalent to the recorded data in the aforementioned embodiments. The recorded data may further include user data, business scope, demand cycle, and business level. In this example, user data may be equivalent to the behavior type information in the aforementioned embodiments; business scope may be equivalent to the usage degree information in the aforementioned embodiments; demand cycle may be equivalent to the preset cycle information in the aforementioned embodiments; and business level may be equivalent to the user-defined information in the aforementioned embodiments.

[0098] User data (UA, User Actions) represents user-related data concerning their daily life and services. User data can include: offline data from the real world, location-based services (LBS) data that is not directly related to conversions, and LBS data that is both online and related to conversions. Offline data, for example, could record a user's visit to a restaurant and can be categorized as "offline". Online LBS data, for example, could record a user's navigation to a restaurant and can be categorized as "online-non-biz". Online LBS data that is related to conversions, for example, could record a user receiving a coupon for a restaurant and can be categorized as "online-biz". User data can be recorded as UA(ua_type), which can be represented as ua_type:(offline, online-non-biz, online-biz). In this example, LBS can refer to Location Based Services, which can include services based on geographic location data. These services are provided by mobile terminals using wireless communication networks (or satellite positioning systems) and spatial databases to obtain the user's geographic location coordinates and integrate them with other information to provide the user with the location-related value-added services they need, such as food delivery services, ride-hailing services, restaurant reservation services, accommodation services, and same-city delivery services.

[0099] In this example, the business scope BS (biz scope) can be used to represent various vertical businesses in local life. This data relies on the LBS platform's business segmentation. In this application example, the business scope can be divided into seven aspects or types: restaurant, hotel, trip, entertainment, travel, life-service, and shopping. That is, the business scope can be represented as: BS:(restaurant, hotel, trip, entertainment, travel, life-service, shopping). In this example, travel can refer to a user's behavior of visiting tourist attractions. Trip can include a user's daily transportation behavior.

[0100] In this example, the demand cycle (DC) represents the user's demand cycle for the business, which can be divided into daily, weekly, and monthly cycles. For example, if the business scope is food, the demand cycle is daily. In this example, DC and BS can have a one-to-one correspondence. The demand cycle can be represented as: DC:(daily, weekly, monthly).

[0101] In this example, the business level (BL) can be a category or definition of user filtering. From a user growth perspective, user filtering can be divided into three levels. The business level can be represented as: BL:(offline, online_interest, online_order). In this example, BL and ua_type have a one-to-one correspondence.

[0102] Here, "offline" can refer to users with offline needs, meaning people who have a need for the service in the real world. For example, people who want to buy food in the real world. "Online interest" can refer to users with online interests, meaning users who show interest in the service on the current business platform. For example, users searching for food on a map application. "Online order" can refer to users who have converted their needs for the service into actual purchases on the current business platform. For example, people ordering food on a map application.

[0103] In this example, the analysis model is equivalent to the business data processing model in the aforementioned embodiments. It can generate output data representing the degree of user demand by integrating and processing user data, business types, and business demand type data. By inputting different businesses and different preset business types, it can generate demand levels corresponding to different preset business types. Therefore, this analysis model can generate user demand analysis data for all categories of life services. Thus, the expression for the output data representing user demand can be: Demand Level = Analysis Model (User Data, Business Type, Business Level, Demand Cycle), i.e., sod = UBPS(UA(ua_type), bs, bl, dc). Here, sod represents the degree of user demand, and UBPS represents the analysis model.

[0104] In this example, data DC can be represented as: dc:(daily,weekly,monthly). Data BL can be represented as: bl:(offline,online_interest,online_order). Data ua_type can be represented as: ua_type:(offline,online-non-biz,online-biz). Data BS can be represented as: bs:(restaurant,hotel,trip,entertainment,travel,life-service,shopping).

[0105] Since BS and DC have a one-to-one correspondence, meaning that the demand cycle is determined when the business type is determined, the sod formula can be simplified to: sod = UBPS(UA(ua_type),bs,bl).

[0106] In practical applications, business stakeholders often want to generate a full business logic (BL) for users, obtaining comprehensive output information across all levels of the BL. This involves analyzing the recorded data to obtain data outputs representing three layers of user needs: offline, online interests, and online conversions. There is a one-to-one correspondence between ua_type and BL, and the sod formula can be simplified to: sod = UBPS(bs). Once the business stakeholders define the business scope (BS), the analysis model can automatically stratify user needs across the three levels: offline, online interests, and online conversions.

[0107] Generally, users' demand for local life services exhibits cyclical and stable characteristics.

[0108] The aforementioned periodicity refers to the clear recurring patterns in user demand for local lifestyle services. For example, the demand for food is measured in days, so the period for this service could be on the daily level. The demand for travel generally follows the characteristic of weekend travel, so the period for this service could be on the weekly level. As for hotel demand, aside from business travelers, most scenarios involve out-of-town leisure travel, concentrated on holidays, so the period for this service could be on the monthly level. Other categories of travel, entertainment, lifestyle services, and shopping demands can all be categorized into daily, weekly, and monthly periodic patterns.

[0109] The aforementioned stability refers to the fact that users' demand for local lifestyle services remains relatively stable at a certain level over a long period, such as a year. This is because the fundamental factor determining the level of user demand is the user's lifestyle, and for most users, their lifestyle remains relatively fixed over a long period. Even if users' demand for services shifts over time, this shift will be a smooth transition rather than a sudden change. For example, the demand for refueling may vary from person to person over a week, two weeks, or a month, but it generally depends on the frequency and mileage of the user's driving. Similarly, the demand for shopping may vary from person to person over a week or a month, but it generally depends on the user's demand for consumables. The same applies to other categories such as hotels, transportation, entertainment, and tourism services, because the ultimate reflection is the user's lifestyle, and lifestyles are generally stable.

[0110] Leveraging the periodicity and stability of users' behavior regarding local life services, and the relationship between users and services, the processing flowchart of the analytical model used in this application example is as follows: Figure 3As shown. Among these: periodicity and stability: reflected in the core algorithm of the FG (Feature Generator) in this application example, and in the modeling process of the Common Ranking Model (CRM). The relationship between users and business: reflected in the mapping between ua_type and specific user behaviors.

[0111] Reference Figure 3 As shown in the example of this application, the process steps of the business data analysis method include steps 301 to 303.

[0112] In step S301, user behavior is stratified according to the definition of ua_type and filtered according to business type. Here, ua_type = "offline" indicates user offline behavior. In this example, the recorded data used can all come from the user's initial location information (GPS). Through the aggregation and analysis of GPS points throughout the day, user behavior trajectory information is quantified, feature transformation is performed, and feature output is generated. The generated features can include local / out-of-town features, en route features, and arrival point features.

[0113] Among them, the local cross-city feature is used to quantify the user's cross-city behavior; the on-the-go feature is used to quantify the user's behavior while driving; and the store arrival feature is used to describe the relationship between the user and the local life service store (it can identify the user's in-store behavior and offline behavior, and can be recorded by the application).

[0114] ua_type="online-non-biz" indicates user behavior that occurs within LBS (Location-Based Services) and is related to basic LBS services, but not to online conversion behavior related to biz_type. This part mainly comes from basic LBS services such as map applications. Data with ua_type="online-non-biz" can include refined search, broad search, click data, navigation, route data, and daily active user data.

[0115] Specifically, refined search, broad search, and click data can include users' searches on LBS and their filtering behavior on the results. Refined search refers to searching for a specific location, such as searching for "Ski Resort B" in city A. Broad search refers to searching for generalized needs, such as searching for "skiing". Clicks refer to the filtering behavior of search results, such as clicking on a POI after searching for "Ski Resort B".

[0116] Navigation and route data can reflect a user's travel needs on LBS (Location-Based Services). For example, before setting off, a user might use LBS "route" queries to assess travel time to "Ski Resort B," and these queries can be recorded as navigation and route data in this application's example. Similarly, navigation behavior during the user's journey to "Ski Resort B" can also be used as navigation and route data in this application's example.

[0117] Daily active users (DAU) data can represent the stickiness between users and location-based services (LBS). In this example, DAU can be short for Daily Active Users, a statistical metric used to reflect the operational status of a website, internet application, or online game. DAU typically counts the number of users who logged in or used a product within a single day (statistical day) (excluding duplicate logins).

[0118] `ua_type = "online-biz"` can represent user behaviors on LBS that are related to conversions in local life services. This part depends on the specific content of the LBS service and the user's understanding of the transaction-related functions of the LBS platform. In the example of this application, it can specifically include user behaviors on certain products of POI online, including browsing, clicking, conversion, as well as basic platform functions such as claiming coupons, making phone calls, and clicking on content such as advertisements.

[0119] The BS data in this application example can be used to filter behaviors related to a certain category of local life services, such as user agents (UAs) on POIs related to "hotels".

[0120] In step S302, the features output by the FG (Features Generator) are calculated. The output data is as follows: based on the periodicity and stability features, the FG extracts the interpretability features of the UA over time, producing periodic and stability features. The values ​​of the periodic features can include: daily, weekly, and monthly features, representing daily, weekly, and monthly demand levels, respectively. The stability features can be trend features representing stability, and their values ​​can include: smooth, up, and down features, representing stable demand, strengthening demand, and weakening demand within the period, respectively.

[0121] In this application example, when step S302 is executed, the time series feature generator, feature clusterer and FG can be used sequentially to process the data in step S301.

[0122] Among them, the sequence generator can process a user's behavior over multiple days into a time-series vector. For example, a user's time-series behavior in cross-city travel over the past 60 days (or other days) can be modeled as a 60-dimensional vector, where 1 represents cross-city travel and 0 represents no cross-city travel.

[0123] A sequence clusterer, also known as a periodicity / stability feature extractor, clusters temporal features (sequences) and sets a periodicity × stability vector as the centroid feature (in this example, the centroid can be a user behavior vector of 60 or other dimensions). The centroid feature represents the feature whose stability is increasing, stable, or decreasing within the corresponding period, and serves as a reference feature for analyzing temporal features. This application example uses a sequence clusterer to find the closest way to the centroid for the sequence vectors generated by the temporal feature generator to achieve feature clustering analysis. The output can represent the stable, decreasing, or increasing trend of user behavior changes within the corresponding period.

[0124] In the example of this application, the corresponding centroid vector design is as follows: Figure 4A As shown. The dimension of the centroid vector can be consistent with the sampling period; for example, if the sampling period is 60 days, the centroid vector can be 60-dimensional. Simultaneously, the number of centroid vectors can be consistent with the preset period of the business; if the preset period is longer than the minimum time unit, the number of centroid vectors is greater than 1. For example, if the minimum time unit is days, such as... Figure 4A As shown, for a business with a preset cycle of one week, there can be 7 centroid vectors.

[0125] Figure 4A The centroid vector shown is represented numerically as follows:

[0126]

[0127] Reference Figure 4A As shown, the centroid vectors representing the decreasing stability of business-related behaviors can be 401-407. Each vector represents a stable behavioral characteristic occurring on the same day (Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday) within a weekly cycle. For example, in a business with a preset weekly cycle, if the characteristics of a user's business-related behaviors on any of the seven days of the week (i.e., Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday) match any centroid vector from 401-407, then the stability of the user's business-related behaviors can be considered to decrease within a week. Figure 4A In the example shown, for a business with a preset cycle of one week, there are seven centroid vectors. If the time-series characteristics obtained from the analysis of the recorded data are... Figure 4A If any vector in the vector is close to the target vector, then the stability of the user's business-related behavior is considered to be stationary.

[0128] Figure 4B The centroid vector represents the downward trend of business-related behaviors over a preset weekly period. When the centroid vector value is small, it can be plotted by magnifying it. A centroid vector magnified 10 times can be represented numerically as follows:

[0129]

[0130] Reference Figure 4B As shown, the centroid vectors representing the decreasing stability of business-related behaviors can be 408-414. Each vector represents a behavioral characteristic with a decreasing trend occurring on the same day (Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday) within a weekly cycle. For example, in a business with a preset weekly cycle, if the characteristics of a user's business-related behaviors on any of the seven days of the week (i.e., Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday) match any centroid vector from 408-414, then the stability of the user's business-related behaviors can be considered to be decreasing within a week. Figure 4B In the example shown, for a business with a preset cycle of one week, there are seven centroid vectors. If the time-series characteristics obtained from the analysis of the recorded data are... Figure 4B If any vector in the vector is close to the target vector, it indicates that the stability of the user's business-related behavior is decreasing.

[0131] Figure 4C-4F The centroid vector represents the upward trend of business-related behaviors over a preset weekly period. When the centroid vector value is small, it can be represented by a larger image. A centroid vector multiplied by 10 can be numerically represented as follows:

[0132]

[0133] Reference Figure 4C-4F As shown in this example, the centroid vector representing the stability of business-related behaviors is an increasing vector, which can include... Figure 4C-4FIn the vectors 415-421, each vector represents a decreasing trend in behavioral characteristics occurring on the same day (Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday) within a weekly cycle. For example, in a business with a preset weekly cycle, if the characteristics of a user's business-related behavior on any of the seven days of the week (i.e., Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, or Sunday) match any centroid vector in 415-421, then the stability of the user's business-related behavior can be considered to decrease within a week. Figure 4C-4F In the example shown, for a business with a preset cycle of one week, there are seven centroid vectors. If the time-series characteristics obtained from the analysis of the recorded data are... Figure 4C-4F If any vector in the vector is close to the target vector, it indicates that the stability of the user's business-related behavior is decreasing.

[0134] In this application example, the features generated by the FG feature generation algorithm are output, and then the clustering results of the FG-output features are produced based on the analysis of the feature clusterer. The FG algorithm can use the cosine value between vectors as the distance between the time-series features obtained from the recorded information and the pointing vector. In this application embodiment, centroid vectors can be set for the upward trend, stable trend, and downward region of business-related behaviors on a daily period, the downward trend, stable trend, and upward trend of business-related behaviors on a weekly period, and the upward trend, stable trend, and downward trend of business-related behaviors on a monthly period. For the feature generator FG, the input data includes: the time-series vector obtained from the recorded data and the preset centroid vector. In this embodiment, when centroid vectors representing upward, stable, and downward trends are set for business-related behaviors with daily, weekly, and monthly periods, the centroid vector for an upward, stable, or downward trend with a daily period is 1×60 dimensions, the centroid vector for an upward, stable, or downward trend with a weekly period is 7×60 dimensions, and the centroid vector for an upward, downward, or stable trend with a monthly period is 30×60 dimensions. As an example, the centroid vector...

[0135]

[0136] According to the above formula, there can be 114 centroid vectors. A loop function can be set: the FeatureGenerator Cluster performs a comparison operation between the time-series features generated from the recorded data and the 114 centroid vectors. The output vector of the Feature Generator (FG) represents the stability of user-related business behaviors over days, weeks, and months. Assuming that among the centroid vectors, the time-series vector p of the recorded data has the shortest distance to smooth and weekly (stationary, weekly), the result is smooth\weekly, indicating that the stability of the user's weekly business-related behaviors is stationary. The output data of FG can be centroid vectors that are close to the time-series features.

[0137] In step S303, a classification algorithm is executed using the Common Ranking Model to identify potential users for each business from different ua_type levels. Since potential users exhibit daily, weekly, and monthly cyclical behavior towards local life services, only general daily, weekly, and monthly models are needed. This approach can solve the user segmentation problem in cold start scenarios. The modeling process and CRM structure are as follows: Figure 5 As shown.

[0138] Reference Figure 5 The input features of the CRM model are the output features of different ua_types corresponding to FG in step S302.

[0139] Still refer to Figure 5 During the CRM model training phase, a set of samples is constructed for relevant behaviors on a daily, weekly, and monthly basis. These samples are obtained by extracting the target user's output features from the output features of S302 and labeling them using a specific method. In the services provided by the map application, the daily samples are derived from frequent food users, the weekly samples from frequent gas station users, and the monthly samples from frequent hotel users.

[0140] Still refer to Figure 5 The input data processing model preceding the CRM model can employ general classification models such as RF (Random Forest) or GBDT (Gradient Boosting Decision Tree). In scenarios involving the analysis of business aspects related to map applications, the RF model can be used. The input data to the RF model can include three types of samples × three sets of FG features. Each set of features for each type of sample corresponds to an input data model used to analyze the CRM input features. The aforementioned three types of samples include: daily samples, weekly samples, and monthly samples. The aforementioned three sets of FG features include: daily-level stability FG features, weekly-level stability FG features, and monthly-level stability FG features. Therefore... Figure 5 In the example shown, the input data analysis model includes nine sub-models. Based on the output features of these nine input data analysis models, three CRM models are aggregated at the ua_type level to produce three general hierarchical model outputs.

[0141] Generally, analyzing user needs requires user information such as age, gender, family information, and company information. However, associating this user information with local life services presents challenges such as high association costs and unpredictable results. The method provided in this application reduces business association costs: businesses can directly acquire potential users; storage and computing resources are reduced: due to the universality of the model used in this application, the storage and computing costs of data used for analyzing user needs are significantly reduced; and cold start is resolved: the analysis model used in this application is a universal model applicable to multiple categories. Therefore, for newly launched businesses, there is no need to build samples or upgrade the model; simply input the business type (BS) and determine the BS demand cycle (DC) (daily, weekly, monthly), and the user's conversion potential across the three business lines can be automatically calculated. Figure 6 As shown, this application embodiment can provide comprehensive demand analysis data about users. Through this application embodiment, a user's demand level information for charging is 0.1; demand level information for green travel (which may include the behavior of using public transportation, cycling, walking, etc.) is 0.0; demand level information for car rental is 0.4; demand level information for family and children is 0.8; demand level information for beauty is 0.6; demand level information for tourism is 0.7; demand level information for food is 0.8; demand level information for hotels is 0.2; demand level information for refueling is 0.4; and demand level information for life services is 0.8.

[0142] In this embodiment of the application, based on the user growth system architecture of map applications in local life and the analysis of user demand information, breakthrough progress has been achieved, such as... Figure 7 As shown, for users with food-related needs generated by the analysis model, more food lists are pushed to them, while users are guided to use the map for navigation, thereby increasing user activity.

[0143] The business data analysis method, recommendation information generation method, and recommendation method provided in this application can automatically associate business processes and reduce business application costs. By associating user behavior (past and incremental behavior) with business scope (BS) and business level (BL), it achieves the output of user demand levels across all categories and levels, establishing a strong connection between business and users. It designs general feature processing and model ranking stages, catering to demand across all categories: through a general, cluster-based feature generation algorithm FG (strengthened, stable, weakened, clustering algorithm, behavior strengthening / weakening), it automatically extracts periodic and stable user behavior data across all categories, serving as interpretability input for the general model; through the application of a general, classification-based demand hierarchical model CRM (Common Ranking Model), it extracts user demand levels across multiple categories. Therefore, this application example achieves low resource consumption and low maintenance costs, requiring only the upgrading and maintenance of one model, CRM (common ranking model). (Example: User demand levels in hotels.)

[0144] Corresponding to the business data analysis device provided in the embodiments of this application, the embodiments of this application also provide a business data processing device, such as... Figure 8 As shown, it includes: a data analysis module 801, used to determine the service type information and user behavior information of the user based on the recorded data of the user's business-related behavior; a related behavior change information module 802, used to obtain the user's related behavior change information on the preset service type within a preset period based on the service type information and the user behavior information; and a demand level information module 803, used to obtain the user's demand level information on the preset service type based on the user's related behavior change information on the preset service type within a preset period.

[0145] In one embodiment of this application, the business data processing apparatus further includes: a data recording module, configured to obtain recorded data of the user's use of the map application as recorded data of the user's business-related behavior when the user uses the application to perform business behaviors related to points of interest in the map application.

[0146] In one embodiment of this application, the data analysis module includes: a usage level unit, used to obtain user usage level information for at least one preset business type based on the recorded data of user business-related behaviors; the at least one preset business type includes a preset business type; and a level processing unit, used to treat the usage level information as business type information.

[0147] In one embodiment of this application, the usage level unit is further configured to: obtain the degree of closeness between the service used by the user and at least one preset service type; and obtain usage level information based on the recorded data and the degree of closeness.

[0148] In one embodiment of this application, the recorded data analysis module includes: a behavior type and user-defined information unit, used to obtain user behavior type information and user-defined information corresponding to the behavior type information based on the recorded data of user business-related behaviors; and a behavior definition processing unit, used to use the behavior type information and user-defined information as usage behavior information.

[0149] In one embodiment of this application, the behavior type and user-defined information unit are further configured to perform at least one of the following: if the recorded data indicates that the user completes a business activity at a physical address, determine that the behavior type information is an offline type; if the recorded data indicates that the user queries a business online but does not complete the business activity online, determine that the behavior type information is an online non-business type; if the recorded data indicates that the user completes a business activity online, determine that the behavior type information is an online business type.

[0150] In one embodiment of this application, the relevant behavior change information module includes: a time-series feature unit, used to analyze the stability of user usage behavior of a preset business type within a sampling period based on business type information and usage behavior information, and generate corresponding time-series features; a distance information unit, used to calculate the distance information between the time-series features and the centroid features; the centroid features are a first centroid feature corresponding to stable business behavior occurring within the preset period, a second centroid feature corresponding to upward trending business behavior occurring within the preset period, and a third centroid feature corresponding to downward trending business behavior occurring within the preset period; and a change information unit, used to determine the relevant behavior change information of the user on the preset business type within the preset period based on the distance information.

[0151] In one embodiment of this application, the change information unit is further configured to: determine the stability of the user's related behavior on a preset service type within a preset period, and use the stability as usage behavior change information, when the time-series feature is most closely related to the first centroid feature among the first centroid feature, second centroid feature, and third centroid feature, based on the distance information; determine the rise of the user's related behavior on a preset service type within a preset period, and use the rise as usage behavior change information, when the time-series feature is most closely related to the second centroid feature among the first centroid feature, second centroid feature, and third centroid feature, based on the distance information; and determine the fall of the user's related behavior on a preset service type within a preset period, and use the fall as usage behavior change information, when the time-series feature is most closely related to the third centroid feature among the first centroid feature, second centroid feature, and third centroid feature, based on the distance information.

[0152] In one embodiment of this application, the sampling period is greater than a preset period.

[0153] In one embodiment of this application, the preset period includes multiple periods. For example, the preset period includes days, weeks, and months.

[0154] This application also provides a method for training a business data processing model, comprising: a data input module for inputting training samples into a business data processing model to be trained, such that the business data processing model to be trained executes the business data processing method provided in any embodiment of this application to obtain an estimated value of demand level information; and a training module for training the business data processing model to be trained based on the reference value of demand level information and the estimated value of demand level information of the training samples to obtain a trained business data processing model.

[0155] This application also provides an information recommendation device, including: a business determination module, used to determine the target business information required by the user based on the user's demand level information for a preset business type; the demand level information is generated by the business data processing device provided in any embodiment of this application; and a recommendation information generation module, used to generate recommendation information based on the target business information required by the user.

[0156] This application also provides an information recommendation device, including: a recommendation information acquisition module, used to acquire recommendation information based on an event that has occurred under the detected recommendation conditions; the recommendation information is generated by the recommendation information generation device provided in any embodiment of this application; and a recommendation information sending module, used to send the recommendation information to a user.

[0157] The functions of each module in each device in the embodiments of this application can be found in the corresponding description in the above method, and they have corresponding beneficial effects, which will not be repeated here.

[0158] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc., such as the recorded data in the aforementioned embodiments) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0159] Figure 9 This is a block diagram of an electronic device used to implement embodiments of this application. For example... Figure 9As shown, the electronic device includes a memory 610 and a processor 620. The memory 610 stores a computer program that can run on the processor 620. When the processor 620 executes the computer program, it implements the methods described in the above embodiments. The number of memories 610 and processors 620 can be one or more.

[0160] The electronic device also includes:

[0161] The communication interface 630 is used to communicate with external devices and perform data exchange and transmission.

[0162] If the memory 610, processor 620, and communication interface 630 are implemented independently, they can be interconnected via a bus to communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0163] Optionally, in a specific implementation, if the memory 610, processor 620, and communication interface 630 are integrated on a single chip, then the memory 610, processor 620, and communication interface 630 can communicate with each other through an internal interface.

[0164] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method provided in this application.

[0165] This application also provides a chip including a processor for calling and executing instructions stored in a memory, causing a communication device with the chip installed to perform the method provided in this application.

[0166] This application also provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, output interface, processor, and memory are connected through an internal connection path. The processor is used to execute code in the memory. When the code is executed, the processor is used to execute the method provided in the application embodiment.

[0167] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting Advanced Reduced Instruction Set Machines (ARM) architecture.

[0168] Further, optionally, the aforementioned memory may include read-only memory and random access memory. The memory may be volatile memory or non-volatile memory, or may include both. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0169] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to this application are generated. 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 transferred from one computer-readable storage medium to another.

[0170] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0171] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0172] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0173] The logic and / or steps described in the flowchart or otherwise herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0174] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, the program being stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiments.

[0175] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. This storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0176] The above description is merely an exemplary embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope described in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A business data processing method, characterized in that, include: Based on the recorded data of users' business-related behaviors, determine the types of services used by users and the users' usage behavior information; Based on the service type information and the usage behavior information, analyze the stability of user usage behavior for the preset service type within the sampling period, and generate corresponding time-series features; Calculate the distance information between the time-series features and the centroid features; the centroid features are the first centroid features corresponding to stable business behaviors occurring within a preset period, the second centroid features corresponding to upward business behaviors occurring within a preset period, and the third centroid features corresponding to downward business behaviors occurring within a preset period. Based on the distance information, determine the user's behavioral changes related to the preset service type within a preset period; Based on the user's behavioral changes related to the preset service type within a preset period, the user's demand for the preset service type is determined.

2. The method according to claim 1, characterized in that, The process of determining the type of service used by the user based on recorded data of the user's business-related behaviors includes: Based on the recorded data of users' business-related behaviors, obtain information on the degree of user usage of at least one preset business type; The usage level information is used as the business type information.

3. The method according to claim 2, characterized in that, Based on the recorded data of users' business-related behaviors, information on the degree of user usage of at least one preset business type is obtained, including: Obtain the degree of closeness between the service used by the user and at least one preset service type; The degree of use information is obtained based on the recorded data and the density.

4. The method according to claim 1, characterized in that, The process of determining user behavior information based on recorded data of user business-related behaviors includes: Based on the recorded data of the user's business-related behaviors, obtain the user's behavior type information and the user definition information corresponding to the behavior type information; The behavior type information and the user-defined information are used as the usage behavior information.

5. The method according to claim 4, characterized in that, The process of obtaining user behavior type information based on recorded data of user business-related behaviors includes at least one of the following: When the recorded data indicates that a user completed a business activity at a physical address, the behavior type information is determined to be of the offline type. If the recorded data indicates that a user inquired about a service online but did not complete the service activity online, the behavior type information is determined to be an online non-service type. If the recorded data indicates that a user has completed a business activity online, the behavior type information is determined to be an online business type.

6. The method according to claim 1, characterized in that, Based on the distance information, determine the user's behavioral changes related to the preset service type within a preset period, including: Based on the distance information, if the temporal feature is most closely related to the first centroid feature among the first centroid feature, the second centroid feature, and the third centroid feature, the stability of the user's related behavior on the preset service type within the preset period is determined, and the stability is used as the usage behavior change information. Based on the distance information, if it is determined that the temporal feature is closest to the second centroid feature among the first centroid feature, the second centroid feature, and the third centroid feature, the degree of increase of the user's related behavior to the preset service type within the preset period is determined, and the degree of increase is used as the usage behavior change information; Based on the distance information, if the temporal feature is most closely related to the third centroid feature among the first centroid feature, the second centroid feature, and the third centroid feature, the degree of decrease of the user's related behavior to the preset service type within a preset period is determined, and the degree of decrease is used as the usage behavior change information.

7. The method according to any one of claims 1-6, characterized in that, The preset period is set to correspond to the preset service type.

8. A method for training a business data processing model, characterized in that, include: The training samples are input into the business data processing model to be trained, so that the business data processing model to be trained executes the method described in any one of claims 1-7 to obtain the demand level information estimate. Based on the reference value of the demand information of the training samples and the estimated value of the demand information, the business data processing model to be trained is trained to obtain the trained business data processing model.

9. An information recommendation method, characterized in that, include: Based on the user's demand for the preset service types, determine the target service information required by the user. The demand level information is generated by the method described in any one of claims 1-7; Based on the target business information, recommendation information is generated.

10. A business data processing device, characterized in that, include: The data analysis module is used to determine the type of service used by the user and the user's usage behavior information based on the recorded data of the user's business-related behavior. The relevant behavior change information module is used to analyze the stability of user's business usage behavior for a preset business type within the sampling period based on the business type information and the usage behavior information, and generate corresponding time-series features; Calculate the distance information between the time-series features and the centroid features; the centroid features are the first centroid features corresponding to stable business behaviors occurring within a preset period, the second centroid features corresponding to upward business behaviors occurring within a preset period, and the third centroid features corresponding to downward business behaviors occurring within a preset period. Based on the distance information, determine the user's behavioral changes related to the preset service type within a preset period; The demand level information module is used to determine the user's demand level information for the preset service type based on the user's behavioral change information related to the preset service type within a preset period.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements the method of any one of claims 1-9.

12. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of any one of claims 1-9.

Images