User stratification method and system

Abstract

The present application relates to a user stratification method, in which business personnel only need to input business scenario data into a business terminal, and the business terminal can automatically perform user stratification operations and obtain user stratification results, without the need for professional data processing personnel to intervene, thereby eliminating the process of communicating with data processing personnel. In addition, the present method unifies and standardizes user static data and user dynamic data, reducing the result errors caused by inconsistent data underlying processing logic.

Description

Technical Field

[0001] The present application relates to the technical field of Internet data processing, and in particular to a user stratification method and system. Background Art

[0002] The user stratification method refers to the process of screening users in a specific business scenario to filter out the target user groups that meet the needs of the business scenario. In this process, user data is needed, and the required user data mainly includes user static data and dynamic data. User static data usually refers to the attribute characteristics of the user itself, such as gender, age, nature of work, etc. User dynamic data is usually the user's behavioral characteristics, such as browsing web pages, purchasing goods, etc., which is a form of dynamically generated data.

[0003] When business personnel make user stratification requests based on business scenarios, in traditional user stratification methods, data processing personnel generally filter out the required user groups through manual review.

[0004] However, this user stratification method will cause inaccurate user stratification results. This is due to asymmetric business information. The nature of the work of business personnel and data processing personnel is different. The communication cost between them is high. In the communication process, business scenario requirements are often misinterpreted, resulting in a mismatch between user stratification results and business scenario requirements, and the failure to meet business scenario requirements. In addition, the user's static and dynamic data are distributed in many database tables. Different data processing personnel have different technical levels and their understanding of database tables is different. The user stratification results obtained are different, resulting in uneven accuracy. Summary of the invention

[0005] Based on this, it is necessary to provide a user stratification method and system to address the problem that the traditional user stratification method is to filter out the required user groups through manual review, which leads to inaccurate user stratification results.

[0006] The present application provides a user stratification method, which is applied to a service terminal. The user stratification method includes:

[0007] Obtain user data of all users once every preset time period;

[0008] Clean the user data of each user to obtain user static data and user dynamic data in a standardized format; the user static data includes the coupling relationship between the user device unique identifier and the static data field, and the user dynamic data includes the coupling relationship between the user device unique identifier and the user behavior data;

[0009] Store each user's static data and user dynamic data in a standardized format into the server;

[0010] When receiving a user hierarchical request, read the business scenario data and input the business scenario data into the hierarchical model;

[0011] Running the hierarchical model to output a unique identifier of a user device that matches the business scenario data;

[0012] The unique identifier of the user device matching the business scenario data is output, and the method of obtaining the user data every preset time period is returned.

[0013] Furthermore, obtaining user data of all users from the server once every preset time period includes:

[0014] Obtaining business result data of all users from the server at preset time intervals; the business result data includes one or more of user commodity purchase records, user registration form filling records, and user complaint records;

[0015] Extract user behavior data of all users in local storage;

[0016] The third-party business data of all users is obtained through a third-party communication interface, wherein the third-party business data includes one or more of the user's account data on the third-party platform, the user's customer service communication text data on the third-party platform, and the user's customer service communication voice data on the third-party platform.

[0017] Furthermore, the user data of each user is cleaned to obtain user static data and user dynamic data in a standardized format, including:

[0018] Select user data of a user and read the unique identifier of the user device in the user data;

[0019] Take out the string format field in the user data as enumeration type static data;

[0020] Take out the int format field in the user data as numeric static data;

[0021] Extract the timestamp field in the user data as time-based static data;

[0022] Create a user static data table, and place enumeration type static data, numerical type static data and time type static data corresponding to each user device unique identifier, and the user device unique identifier into the user static data table;

[0023] The user data of the selected user is returned, and the unique identifier of the user device in the user data is read until the user data of each user is cleaned.

[0024] Further, before returning the user data of the selected user and reading the unique identifier of the user device in the user data, until the user data of each user is cleaned, the user data of each user is cleaned to obtain the user static data and user dynamic data in a standardized format, further comprising:

[0025] Select user data of a user and read the unique identifier of the user device in the user data;

[0026] Extracting at least one piece of user behavior data from the user data;

[0027] Select a piece of user behavior data;

[0028] Converting the piece of user behavior data into a behavior event ID and multiple behavior event parameters associated with the behavior event ID;

[0029] Returning to the selected piece of user behavior data until each piece of user behavior data has been converted;

[0030] Return the user data of the selected user until all user behavior data in the user data of each user has been converted;

[0031] A user dynamic data table is established, and a behavior event ID corresponding to each user device unique identifier, and multiple behavior event parameters associated with the behavior event ID and the user device unique identifier are placed in the user dynamic data table in correspondence.

[0032] Furthermore, the behavior event parameters include one or more of enumerated behavior event parameters, numerical behavior event parameters and time behavior event parameters.

[0033] Further, the running of the hierarchical model to output a unique identifier of a user device matching the service scenario data includes:

[0034] Run the hierarchical model, and control the condition extraction module in the hierarchical model to extract at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field, and a logical preposition configuration item from the business scenario data;

[0035] Controlling the syntax parsing module in the hierarchical model to combine at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field and the logical preposition configuration item and converting them into an SQL query statement;

[0036] Matching the SQL query statement with the user static data and user dynamic data in standardized formats of all users in the server respectively, and obtaining the unique identifier of the user device that hits the SQL query statement;

[0037] The unique identifier of the user device that hits the SQL query statement is used as the unique identifier of the user device that matches the business scenario data.

[0038] Furthermore, the static preposition configuration items include one or more of greater than, less than and equal to; the dynamic preposition configuration items include one or more of greater than, less than and equal to; and the logical preposition configuration items include one of and and or.

[0039] Further, the grammar parsing module in the controlled hierarchical model combines at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field and the logical preposition configuration item and converts them into an SQL query statement, including:

[0040] Combine at least one static filter condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic filter condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition to generate a JSON statement;

[0041] Read the identifier in the JSON statement, and split the parent condition and child condition according to the identifier in the JSON statement;

[0042] Read the conditional keywords in the JSON statement and convert them into SQL keywords;

[0043] Combine all SQL keywords into a SQL query statement.

[0044] Furthermore, the SQL query statement is matched with the user static data and user dynamic data in a standardized format of all users in the server to obtain the unique identifier of the user device that hits the SQL query statement, including:

[0045] Retrieving a user static data table from a server, matching an SQL query statement with the user static data table, obtaining a user device unique identifier in the user static data table that matches the SQL query statement, and adding the user device unique identifier in the user static data table that matches the SQL query statement into a first unique identifier set;

[0046] Retrieving a user dynamic data table from a server, matching an SQL query statement with the user dynamic data table, obtaining a user device unique identifier in the user dynamic data table that matches the SQL query statement, and adding the user device unique identifier in the user dynamic data table that matches the SQL query statement into a second unique identifier set;

[0047] Taking the intersection of the first unique identifier set and the second unique identifier set to obtain a third unique identifier set;

[0048] All user equipment unique identifiers in the third unique identifier set are used as user equipment unique identifiers that match the SQL query statement.

[0049] The present application provides a user stratification system, comprising:

[0050] A service terminal, used to execute the user stratification method mentioned above; the service terminal includes a memory;

[0051] The server is communicatively connected with the service terminal.

[0052] The present application relates to a user stratification method, in which business personnel only need to input business scenario data into a business terminal, and the business terminal can automatically perform user stratification operations and obtain user stratification results, without the need for professional data processing personnel to intervene, thereby eliminating the process of communicating with data processing personnel. In addition, the present method unifies and standardizes user static data and user dynamic data, reducing the result errors caused by inconsistent data underlying processing logic. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flowchart of a user stratification method provided in an embodiment of the present application.

[0054] Figure 2 A schematic diagram of the structure of a user stratification system provided in one embodiment of the present application.

[0055] Reference numerals:

[0056] 100-service terminal; 110-storage; 200-server. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of this application more clear, the following is a further detailed description of this application in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0058] The present application provides a user stratification method and system.

[0059] On the one hand, the present application provides a user stratification method. It should be noted that the user stratification method provided by the present application is applied to a business terminal. That is, the user stratification method provided by the present application does not require professional data processing personnel to operate, but only business personnel to operate.

[0060] In addition, the user stratification method provided in the present application does not limit its execution subject. Optionally, the execution subject of the user stratification method provided in the present application can be a service terminal. Specifically, the execution subject of the user stratification method provided in the present application can be one or more processors in the service terminal.

[0061] like Figure 1 As shown, in one embodiment of the present application, the user stratification method is applied to a service terminal.

[0062] The user stratification method comprises:

[0063] S100: Acquire user data of all users once every preset time period.

[0064] Specifically, when a user registers an account, browses information on an application, or places an order for a purchase, user data will be generated, and the server will promptly collect and store the user data in the server. Optionally, some user data is stored in a local storage, and some user data is stored in a server. The service terminal will periodically obtain user data of all users from the server and local storage at one time.

[0065] Optionally, the preset time period may be 12 hours. Optionally, the preset time period may be 24 hours.

[0066] S200, clean the user data of each user to obtain user static data and user dynamic data in a standardized format. The user static data includes the coupling relationship between the user device unique identifier and the static data field. The user dynamic data includes the coupling relationship between the user device unique identifier and the user behavior data.

[0067] Specifically, user data has different sources and formats. This step can integrate user data from different sources and in different formats, and clean the user data uniformly according to a pre-set format. The main purpose is to form a standardized user data format.

[0068] In addition, this step can also distinguish user static data from user dynamic data to facilitate subsequent processing.

[0069] S300, storing the user static data and user dynamic data of each user in a standardized format in a server.

[0070] Specifically, the server may set up two different storage areas to store user static data and user dynamic data respectively.

[0071] S400, when receiving a user hierarchical request, reading business scenario data, and inputting the business scenario data into a hierarchical model.

[0072] Specifically, S100 to S300 are the process of regularly cleaning data, and S400 to S600 are the specific hierarchical process.

[0073] Optionally, when the service terminal receives a user stratification request, the service terminal captures the business scenario data included in the user stratification request and inputs the business scenario data into a stratification model in the service terminal.

[0074] S500, running the hierarchical model, and outputting a unique identifier of a user device that matches the business scenario data.

[0075] Specifically, the hierarchical model is a trained deep learning model that can automatically filter the unique identification output of user devices that match the business scenario data without the intervention of professional data processing personnel.

[0076] S600, outputting the unique identifier of the user equipment matching the business scenario data, and returning to S100.

[0077] Specifically, the ultimate goal of this method is to filter out unique identifiers of user devices that meet the needs of business scenarios.

[0078] In this embodiment, the business personnel only need to input the business scenario data into the business terminal, and the business terminal can automatically perform the user stratification operation and obtain the user stratification result without the intervention of professional data processing personnel, thus eliminating the process of communicating with data processing personnel. In addition, this method unifies and standardizes the user static data and user dynamic data, reducing the result error caused by inconsistent data underlying processing logic.

[0079] In one embodiment of the present application, the S100 includes:

[0080] S110, obtaining business result data of all users from the server at every preset time period. The business result data includes one or more of user commodity purchase records, user registration form filling records and user complaint records.

[0081] S120, extracting user behavior data of all users from the local storage.

[0082] S130, obtaining third-party business data of all users through a third-party communication interface, wherein the third-party business data includes one or more of the user's account data on the third-party platform, the user's customer service communication text data on the third-party platform, and the user's customer service communication voice data on the third-party platform.

[0083] Specifically, user data includes three types of data: business result data, user behavior data and third-party business data. The business result data includes one or more of user commodity purchase records, user registration form filling records and user complaint records.

[0084] User behavior data is generally buried data, collected by business terminals. User behavior data records user behaviors such as opening pages and browsing pages, such as: the time when the behavior occurred (2022-03-07 10:00:00), the name of the web page (XX web page), the user device ID (which can be the IMEI code, such as 123456), the length of stay (5 seconds), etc.

[0085] Third-party business data is data provided by a third-party platform through a third-party communication interface. A third-party platform refers to a platform other than a user terminal and a business terminal. For example, if a chat software platform developed by Enterprise A has signed an information interaction agreement with a business terminal, the account data of the user registered on the chat software platform developed by Enterprise A, and the text data and voice data of the user communicating with customers on the chat software platform developed by Enterprise A will be captured by the business terminal through the third-party communication interface.

[0086] In one embodiment of the present application, the S200 includes:

[0087] S211, selecting user data of a user, and reading a unique identifier of a user device in the user data.

[0088] S212, taking out the string format field in the user data as enumeration type static data.

[0089] S213, extract the int format field in the user data as numeric static data.

[0090] S214, extract the timestamp field in the user data as time-type static data.

[0091] S215, creating a user static data table, and placing enumeration type static data, numerical type static data and time type static data corresponding to each user equipment unique identifier, and the user equipment unique identifier into the user static data table.

[0092] S230, returning to S211, until the user data of each user is cleaned.

[0093] Specifically, this embodiment mainly introduces the format unification of user static data. We define the string format field in the user data as enumerated static data. This type of data indicates the personal characteristics of the user and rarely changes. For example: gender is male, place of origin is Beijing, and occupation type is civil servant.

[0094] We define the int format fields in user data as numeric static data. This type of data indicates some measurable characteristics of the user, which will change over time. For example: age is 26, and the cumulative consumption amount is 1000.

[0095] We define the timestamp field in user data as time-based static data. This type of data indicates the characteristics of the user at a certain point in time. For example, the birthday is 1979-02-04 and the last consumption time is 2020-06-07.

[0096] In order to make the user static data in the standardized format finally formed include the coupling relationship between the user device (user) and the user static data, in S215, we establish a user static data table and use the user static data table as the user static data in the standardized format.

[0097] Table 1 - User static data table (exemplary)

[0098]

[0099] Table 1 is an implementation of the user static data table. Each user has a unique user ID. The unique user ID can be numbered by the service terminal, or the native device unique ID IMEI code of the user terminal can be used. Gender is enumerated static data. The last consumption time is time-based static data. Age and cumulative consumption amount are numerical static data.

[0100] The result recorded in each row is a static label on the user. It is worth mentioning that since there is only one row of results for a user's static data, the user's static data is also unique, and it is impossible to record the static data change process. Therefore, enumeration-type static data can only record the latest result, time-type enumeration-type static data can only record the first and most recent results, and numerical enumeration-type static data can only record counts and sums. This has certain limitations in the application process, so user dynamic data is needed to supplement it to make up for this deficiency.

[0101] In one embodiment of the present application, before S230, S200 further includes:

[0102] S221, selecting user data of a user, and reading a unique identifier of a user device in the user data.

[0103] Specifically, each user has a unique user ID, which can be numbered by a service terminal or can be a native device unique ID IMEI code of the user terminal.

[0104] S222: extract at least one piece of user behavior data from the user data.

[0105] S223, select a piece of user behavior data.

[0106] S224: Convert the piece of user behavior data into a behavior event ID and a plurality of behavior event parameters associated with the behavior event ID.

[0107] S225, returning to S223, until each piece of user behavior data is converted.

[0108] Specifically, S223 to S224 are repeatedly executed until each piece of user behavior data is converted.

[0109] S226, returning to S221, until all user behavior data in the user data of each user is converted.

[0110] Specifically, S222 to S225 are repeatedly performed until all user behavior data in the user data of each user are converted.

[0111] S227, establishing a user dynamic data table, and placing the behavior event ID corresponding to each user device unique identifier, and multiple behavior event parameters associated with the behavior event ID and the user device unique identifier into the user dynamic data table in correspondence.

[0112] Specifically, in order to make the user dynamic data in the standardized format finally formed contain the coupling relationship between the user device (user) and the user dynamic data, we established a user dynamic data table and used the user dynamic data table as the user dynamic data in the standardized format.

[0113] Table 2 - User dynamic data table (exemplary)

[0114]

[0115] Table 2 is an implementation of the user dynamic data table.

[0116] As shown in Table 2, each line in the table is a piece of user behavior data for a user. The behavior event IDs are K001 and K002, K001 represents the behavior event of purchasing a product, and K002 represents the behavior event of browsing a page.

[0117] In one embodiment of the present application, the behavior event parameters include one or more of enumerated behavior event parameters, numerical behavior event parameters, and time behavior event parameters.

[0118] Specifically, user behavior data also has different types, which are represented by different types of behavior event parameters. The behavior event parameters are further specific descriptions of the behavior events.

[0119] As shown in Table 2, the enumeration-type behavior event parameter in Table 2 represents the price of the purchased product. The time-type behavior event parameter in Table 2 represents the time when the user behavior event occurs. The numerical behavior event parameter in Table 2 represents the number of purchased products in the behavior event K001, and represents the ID of the page browsed by the user in the behavior event K002 (i.e., Page_1 and Page_2).

[0120] Since browsing a page does not generate the price of the purchased product, its enumerated behavior event parameter is represented by "-".

[0121] In one embodiment of the present application, S500 includes:

[0122] S510, run the hierarchical model, and control the conditional extraction module in the hierarchical model to extract at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field, and a logical preposition configuration item from the business scenario data.

[0123] S520, controlling the syntax parsing module in the hierarchical model to merge at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition configuration item and convert them into an SQL query statement.

[0124] S530, matching the SQL query statement with the user static data and user dynamic data in standardized formats of all users in the server respectively, and obtaining the unique identifier of the user device that matches the SQL query statement.

[0125] S540: Use the unique identifier of the user device that matches the SQL query statement as the unique identifier of the user device that matches the business scenario data.

[0126] Specifically, the business scenario data includes business scenario requirement information, from which the business terminal can extract at least one static filtering condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic filtering condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition configuration item.

[0127] For example, the business scenario demand information included in the business scenario data is: women who purchase more than one item between December 20, 2021 and December 30, 2022, and the price of the purchased items is greater than 100.

[0128] Then the conditional extraction module can extract only one static filter condition field from the business scenario demand information, which is gender. There is only one static preposition configuration item, which is equal to. There is only one static content field, which is female. There are three dynamic filter condition fields, one is between December 20, 2021 and December 30, 2022, one is purchasing goods, and the other is the price of the goods. There are two dynamic preposition configuration items, both of which are greater than. There are two dynamic content fields, one is 1 and the other is 100. The logical preposition configuration item is and.

[0129] The syntax parsing module can merge at least one static filtering condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic filtering condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition configuration item without destroying the original meaning of the business scenario requirement information, and convert them into an SQL query statement.

[0130] Optionally, when the syntax parsing module performs the merge in S520, it first merges at least one static filter condition field, at least one static preposition configuration item and at least one static content field to obtain a static merge result. Continuing with the above example, the static merge result is "gender equals female". The syntax parsing module further merges at least one dynamic filter condition field, at least one dynamic preposition configuration item and at least one dynamic content field to obtain a dynamic merge result. Continuing with the above example, the dynamic merge result is "purchase of goods is greater than 1, and the price of the goods is greater than 100". The syntax parsing module finally concatenates the static merge result and the dynamic merge result through the logical preposition configuration item to obtain the final merged result, and then converts the final merged result into an SQL query statement.

[0131] In one embodiment of the present application, the static preposition configuration items include one or more of greater than, less than and equal to; the dynamic preposition configuration items include one or more of greater than, less than and equal to; the logical preposition configuration items include one of and and or.

[0132] Specifically, it can be understood that the static filtering condition fields also have enumeration type, time type and numeric type, which are completely corresponding to the enumeration type static data, numeric type static data and time type static data in the standardized format of user static data. The dynamic filtering condition fields also have enumeration type, time type and numeric type, which are completely corresponding to the enumeration type behavior event parameters, numeric type behavior event parameters and time type behavior event parameters in the standardized format of user dynamic data. This is also the reason why S530 can match.

[0133] In one embodiment of the present application, the S520 includes:

[0134] S521, at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition combination are concatenated and a JSON statement is generated.

[0135] For example, the business scenario demand information included in the business scenario data is: women who purchased more than one item between December 20, 2021 and December 30, 2022, and the price of the purchased items was greater than 100. The resulting JSON statement is as follows:

[0136]

[0137]

[0138] S522, read the identifier in the JSON statement, and split the parent condition and the child condition according to the identifier in the JSON statement.

[0139] Specifically, there are some identifiers in the above JSON statement, for example, "child" is a parent condition identifier, representing a parent condition that is subordinate to the previous level. This step separates the parent condition and the child condition according to the identifiers in the JSON statement.

[0140] S523, reading conditional keywords in the JSON statement, and converting the conditional keywords in the JSON statement into SQL keywords.

[0141] Specifically, for example, if gender is equal to female, and the SQL keyword corresponding to gender is gender, then the converted SQL statement is gender='female'.

[0142] S524, all SQL keyword combinations are concatenated into a SQL query statement.

[0143] Specifically, the concatenated SQL query statement is:

[0144] WITH groupConditionJson AS (

[0146] SELECT user_id

[0147] FROM label_table

[0148] WHERE gender='female'

[0149] ),

[0150] doConditionJson AS (

[0152] SELECT user_id

[0153] ,count(*)

[0154] FROM event_table

[0155] WHERE price>100

[0156] AND evnt_time BETWEEN'2021-12-20'

[0157] AND '2021-12-30'

[0158] GROUP BY user_id

[0159] HAVING count(*)>1 )

[0161] SELECT user_id

[0162] FROM groupConditionJson

[0163] JOIN doConditionJson

[0164] ON groupConditionJson.user_id=doConditionJson.user_id.

[0165] In one embodiment of the present application, the S530 includes:

[0166] S531, retrieve the user static data table from the server, match the SQL query statement with the user static data table, obtain the user device unique identifier in the user static data table that matches the SQL query statement, and add the user device unique identifier in the user static data table that matches the SQL query statement into the first unique identifier set.

[0167] S532, retrieve the user dynamic data table from the server, match the SQL query statement with the user dynamic data table, obtain the user device unique identifier in the user dynamic data table that matches the SQL query statement, and add the user device unique identifier in the user dynamic data table that matches the SQL query statement into the second unique identifier set.

[0168] S533: Take the intersection of the first unique identifier set and the second unique identifier set to obtain a third unique identifier set.

[0169] S534: Use all user equipment unique identifiers in the third unique identifier set as user equipment unique identifiers that match the SQL query statement.

[0170] Specifically, this step is a table lookup process. Note that the user static data table and the user dynamic data table are queried separately first, the query results are obtained separately, and finally the query results are intersected. Finally, the user device unique identifier in the intersection is used as the user device unique identifier that hits the SQL query statement for subsequent use by the business personnel. In other words, the user device unique identifier in the user static data table that matches the SQL query statement may be multiple or one. The user device unique identifier in the user dynamic data table that matches the SQL query statement may be multiple or one.

[0171] On the other hand, the present application also provides a user stratification system.

[0172] like Figure 2 As shown, in one embodiment of the present application, the user stratification system includes a service terminal 100 and a server 200 .

[0173] The service terminal 100 is used to execute the user stratification method mentioned above. The service terminal 100 includes a memory 110. The server 200 is in communication connection with the service terminal 100.

[0174] Specifically, the service terminal 100 is equipped with a hierarchical model, which is a pre-trained deep learning model.

[0175] It should be noted that, for the sake of brevity, the service terminal 100, the storage 110 and the server 200 are only numbered in this embodiment, and in each embodiment of the aforementioned user stratification method, the service terminal 100, the storage 110 and the server 200 are not numbered.

[0176] The technical features of the above-described embodiments may be arbitrarily combined, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0177] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A user stratification method, characterized in that: Applied to a service terminal, the user stratification method includes: Obtain user data of all users once every preset time period; Clean the user data of each user to obtain the user static data and user dynamic data in a standardized format, select the user data of a user, and read the unique identifier of the user device in the user data; Extracting at least one piece of user behavior data from the user data; Select a piece of user behavior data; Converting the piece of user behavior data into a behavior event ID and multiple behavior event parameters associated with the behavior event ID; Returning to the selected piece of user behavior data until each piece of user behavior data has been converted; Return the user data of the selected user until all user behavior data in the user data of each user has been converted; Establish a user dynamic data table, and place the behavior event ID corresponding to each user device unique identifier, and multiple behavior event parameters associated with the behavior event ID and the user device unique identifier in the user dynamic data table in correspondence; the user static data includes the coupling relationship between the user device unique identifier and the static data field, and the user dynamic data includes the coupling relationship between the user device unique identifier and the user behavior data; Store each user's static data and user dynamic data in a standardized format into the server; When receiving a user hierarchical request, read the business scenario data and input the business scenario data into the hierarchical model; Run the hierarchical model, output a unique identifier of a user device that matches the business scenario data, run the hierarchical model, and control a condition extraction module in the hierarchical model to extract at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field, and a logical preposition configuration item from the business scenario data; Controlling the syntax parsing module in the hierarchical model to combine at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field and the logical preposition configuration item and converting them into an SQL query statement; Matching the SQL query statement with the user static data and user dynamic data in standardized formats of all users in the server respectively, and obtaining the unique identifier of the user device that hits the SQL query statement; The unique identifier of the user device that hits the SQL query statement is used as the unique identifier of the user device that matches the business scenario data; The unique identifier of the user device matching the business scenario data is output, and the method of obtaining the user data every preset time period is returned.

2. The user stratification method according to claim 1, characterized in that: The acquiring user data of all users from the server once every preset time period includes: Obtaining business result data of all users from the server at preset time intervals; the business result data includes one or more of user commodity purchase records, user registration form filling records, and user complaint records; Extract user behavior data of all users in local storage; The third-party business data of all users is obtained through a third-party communication interface, wherein the third-party business data includes one or more of the user's account data on the third-party platform, the user's customer service communication text data on the third-party platform, and the user's customer service communication voice data on the third-party platform.

3. The user stratification method according to claim 2, characterized in that: The user data of each user is cleaned to obtain user static data and user dynamic data in a standardized format, including: Select user data of a user and read the unique identifier of the user device in the user data; Take out the string format field in the user data as enumeration type static data; Take out the int format field in the user data as numeric static data; Extract the timestamp field in the user data as time-based static data; Create a user static data table, and place enumeration type static data, numerical type static data and time type static data corresponding to each user device unique identifier, and the user device unique identifier into the user static data table; The user data of the selected user is returned, and the unique identifier of the user device in the user data is read until the user data of each user is cleaned.

4. The user stratification method according to claim 3, characterized in that: The behavior event parameters include one or more of enumeration-type behavior event parameters, numerical-type behavior event parameters, and time-type behavior event parameters.

5. The user stratification method according to claim 4, characterized in that: The static preposition configuration items include one or more of greater than, less than and equal to; the dynamic preposition configuration items include one or more of greater than, less than and equal to; the logical preposition configuration items include one of and and or.

6. The user stratification method according to claim 5, characterized in that: The grammar parsing module in the controlled hierarchical model combines at least one static screening condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic screening condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition configuration item and converts them into an SQL query statement, including: Combine at least one static filter condition field, at least one static preposition configuration item, at least one static content field, at least one dynamic filter condition field, at least one dynamic preposition configuration item, at least one dynamic content field and a logical preposition to generate a JSON statement; Read the identifier in the JSON statement, and split the parent condition and child condition according to the identifier in the JSON statement; Read the conditional keywords in the JSON statement and convert them into SQL keywords; Combine all SQL keywords into a SQL query statement.

7. The user stratification method according to claim 6, characterized in that: The step of matching the SQL query statement with the user static data and user dynamic data in a standardized format of all users in the server to obtain a unique identifier of a user device that hits the SQL query statement includes: Retrieving a user static data table from a server, matching an SQL query statement with the user static data table, obtaining a user device unique identifier in the user static data table that matches the SQL query statement, and adding the user device unique identifier in the user static data table that matches the SQL query statement into a first unique identifier set; Retrieving a user dynamic data table from a server, matching an SQL query statement with the user dynamic data table, obtaining a user device unique identifier in the user dynamic data table that matches the SQL query statement, and adding the user device unique identifier in the user dynamic data table that matches the SQL query statement into a second unique identifier set; Taking the intersection of the first unique identifier set and the second unique identifier set to obtain a third unique identifier set; All user equipment unique identifiers in the third unique identifier set are used as user equipment unique identifiers that match the SQL query statement.

8. A user stratification system, characterized in that: include: A service terminal, configured to execute the user stratification method according to any one of claims 1 to 7; The service terminal includes a memory; The server is communicatively connected with the service terminal.

Images