User identification method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a user recognition method and device, electronic equipment and a storage medium, belongs to the technical field of artificial intelligence, and can improve the recognition accuracy of a maternal and infant customer group. Comprising the steps that a first candidate user group of mother and baby users and information data of the first candidate user group are acquired, the information data comprise internet log data and user basic data, and a first customer group is determined based on the information data; acquiring a second candidate user group except the first customer group in the first candidate user group; inputting the spatio-temporal data of the second candidate user group into the intelligent large model to obtain a user feature vector of the second candidate user group, and inputting the user feature vector into the mother and infant user identification model to obtain a second customer group; the spatio-temporal data comprises data of time and spatial dimensions of a second candidate user group, and the maternal and infant user identification model is used for determining a confidence coefficient that a first user corresponding to the user feature vector is a maternal and infant user; and determining the first customer group and the second customer group as the maternal and infant customer group.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly to a method, device, electronic device and storage medium for user identification. Background Art

[0002] With the development of the digital transformation of the pharmaceutical industry, relevant profit-oriented retail pharmacies and maternal and infant care institutions such as maternity centers have an urgent need to accurately identify the maternal and infant user group. For non-profit units such as maternity hospitals and community health service centers that serve the maternal and infant user group, it is also necessary to accurately identify the maternal and infant user group to empower services such as health consultation, and improve the service satisfaction for maternal and infant users.

[0003] In the related art, generally, the frequency of a user accessing maternal and infant websites, small programs, etc., and the basic user information such as the user's age and permanent residence are used as reference indicators to determine whether the user is a maternal and infant user. However, the penetration rate of the mobile Internet for maternal and infant users in the maternal and infant user group is very low. The target user group mined through the related technical solutions has a low coverage rate of real maternal and infant users. Only considering two dimensions of the user's online log and basic user information results in a low accuracy of identifying the maternal and infant user customer group. Summary of the Invention

[0004] The purpose of the embodiments of this application is to provide a method, device, electronic device and storage medium for user identification, so as to solve the problem of low accuracy in identifying the maternal and infant user customer group.

[0005] To solve the above technical problems, the embodiments of this application are implemented as follows: In a first aspect, the embodiments of this application provide a method for user identification, including: obtaining a first candidate user group of maternal and infant users and information data of the first candidate user group, where the information data includes online log data and basic user data, and determining a first customer group based on the information data; obtaining a second candidate user group in the first candidate user group except the first customer group; inputting the spatio-temporal data of the second candidate user group into an intelligent large model to obtain a user feature vector of the second candidate user group, and inputting the user feature vector into a maternal and infant user identification model to obtain a second customer group; the spatio-temporal data includes data of the second candidate user group in the time and space dimensions, and the maternal and infant user identification model is used to determine the confidence that a first user corresponding to the user feature vector is a maternal and infant user; determining the first customer group and the second customer group as the maternal and infant user customer group.

[0006] In a second aspect, an embodiment of the present application provides a user identification device, including: a first acquisition module, configured to acquire a first candidate user group of the mother and baby users and information data of the first candidate user group, where the information data includes Internet access log data and user basic data, and determine a first customer group based on the information data; a second acquisition module, configured to acquire a second candidate user group other than the first customer group in the first candidate user group; an identification module, configured to input spatio-temporal data of the second candidate user group into an intelligent large model to obtain a user feature vector of the second candidate user group, and input the user feature vector into a mother and baby user identification model to obtain a second customer group; the spatio-temporal data includes data of the second candidate user group in the time and space dimensions, and the mother and baby user identification model is used to determine the confidence that the first user corresponding to the user feature vector is the mother and baby user; a determination module, configured to determine the first customer group and the second customer group as the mother and baby user customer group.

[0007] In a third aspect, an embodiment of the present application provides an electronic device, including a processor and a memory electrically connected to the processor, where the memory stores a computer program, and the processor is configured to call and execute the computer program from the memory to implement the above-mentioned user identification method.

[0008] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium for storing a computer program, and the computer program can be executed by a processor to implement the above-mentioned user identification method.

[0009] In a fifth aspect, an embodiment of the present application provides a chip, where the chip includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is configured to run a program or an instruction to implement the above-mentioned user identification method.

[0010] In a sixth aspect, an embodiment of the present application provides a computer program product, where the computer program product includes a computer program, and when the computer program is executed by a processor, it implements the above-mentioned user identification method.

[0011] Adopt the technical solution of the embodiment of the present application, obtain the first candidate user group of mother and baby users and the information data of the first candidate user group, the information data includes Internet access log data and user basic data, and based on the information data, determine the first customer group; obtain the second candidate user group except the first customer group in the first candidate user group; input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the mother and baby user identification model to obtain the second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the mother and baby user identification model is used to determine the confidence level that the first user corresponding to the user feature vector is a mother and baby user; determine the first customer group and the second customer group as the mother and baby user customer group. It can be seen that through the information data of the first candidate user group of mother and baby users, the first customer group that is a mother and baby user is screened out, and then spatio-temporal data is obtained from the remaining users of the first candidate user group and input into the intelligent large model to obtain the user feature vector. By inputting the user feature vector into the mother and baby user identification model, the second customer group that is a mother and baby user among the remaining users of the first candidate user group is screened out. Through two screening methods, the first customer group and the second customer group are determined as the mother and baby user customer group, which can solve the problem of only considering two dimensions of user Internet access logs and user basic information, make full use of multi-dimensional data and capabilities such as intelligent large models and mother and baby user identification models, and solve the problem of low accuracy in identifying mother and baby user customer groups. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in one or more embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in one or more embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0013] Figure 1 is a schematic flowchart of a user identification method according to an embodiment of the present application; Figure 2 is a schematic diagram of the Chuanliu large model using the mother and baby user identification model to identify mother and baby users according to an embodiment of the present application; Figure 3 is a schematic flowchart of a scenario of a user identification method according to another embodiment of the present application; Figure 4 is a schematic block diagram of a user identification device according to an embodiment of the present application; Figure 5 is a schematic diagram of the hardware structure of a user identification device according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] The embodiments of the present application provide a method, apparatus, electronic device and storage medium for user identification, so as to solve the problem of low accuracy in identifying the customer group of mother and baby users.

[0015] In order to enable those skilled in the art to better understand the technical solutions in the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0016] The method for user identification provided by the embodiments of the present application can be executed by an electronic device or by software installed in the electronic device. Specifically, the electronic device can be a terminal device or a server device. Among them, the terminal device can include a smart phone, a laptop computer, a smart wearable device, a vehicle-mounted terminal, etc., and the server device can include an independent physical server, a server cluster composed of multiple servers, or a cloud server capable of performing cloud computing.

[0017] The following will, with reference to the accompanying drawings, describe in detail a method for user identification provided by the embodiments of the present application through specific embodiments and their application scenarios.

[0018] Figure 1 The following shows a schematic flowchart of a method for user identification provided by an embodiment of the present invention. The method includes the following steps: S102, obtain the first candidate user group of mother and baby users and the information data of the first candidate user group. The information data includes Internet access log data and user basic data, and determine the first customer group based on the information data.

[0019] The first candidate user group refers to the mother and baby users determined through location data in big data. The set of such mother and baby users is the first candidate user group. Among them, the frequency, time, etc. of the user going to the hospital are determined through location data, and then it is judged whether the user is a candidate for mother and baby users. The set of candidates for mother and baby users is determined as the first candidate user group.

[0020] The Internet access log data refers to the Internet access information of the user on mother and baby-related software, websites, mini-programs, etc. within a preset time period. The Internet access information includes, for example, the frequency and time of browsing mother and baby-related software.

[0021] The user basic data refers to the basic information such as the user age, location, geographical location, etc. used to determine each user in the first candidate user group.

[0022] ]>Based on the information data, eligible users in the first candidate user group are determined as the first customer group. Specifically, the information data includes: online log data and basic user data. Within a preset time period, the number of times users in the first candidate user group access maternity and baby-related software, websites, mini-programs, etc. is obtained. A first condition is that the number of times a user accesses maternity and baby-related software is greater than a preset number. Furthermore, the age of users in the first candidate user group is obtained. A second condition is that the user's age is within a preset age range. The set of users that meet both the first and second conditions constitutes the first customer group.

[0023] The first condition is recorded as , the second condition is recorded as , the first customer group is recorded as , the first customer group can be expressed as Formula 1, as shown below: Formula 1 The rule set is { , },function Apply the rule set to the first candidate user group S of suspected mother and baby users That is, the rule set includes the first condition and the second condition, S is the first candidate user group, r is the user who meets both the first condition and the second condition in the first candidate user group, and is selected as the first customer group. .

[0024] As an example, the first condition includes: in the past three months, the total number of visits to maternal and child software, websites, and mini-programs by users in the first candidate user group is greater than 12 times; the second condition includes: the age of users in the first candidate user group is between 20 and 45 years old. Then, the first customer group includes: users in the first candidate user group who meet both of the above conditions.

[0025] S104: Acquire a second candidate user group in the first candidate user group except the first customer group.

[0026] In S102 , a first customer group in the first candidate user group is determined, and a set of users in the first candidate user group excluding the first customer group is determined as a second candidate user group.

[0027] Specifically, the second candidate user group is the complement of the first customer group relative to the first candidate user group. The second candidate user group is recorded as , the first customer group is , the first candidate user group is S, where is the complement symbol, for Relative to the complement of S, the second candidate user group can be expressed as Formula 2, as shown below: Formula 2 S106. Input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the mother and baby user recognition model to obtain the second customer group.

[0028] The spatio-temporal data includes the data of the second candidate user group in the time and space dimensions.

[0029] The mother and baby user recognition model is used to determine the confidence level that the first user corresponding to the user feature vector is a mother and baby user.

[0030] The intelligent large model includes a spatio-temporal large model, which is used to process and analyze spatio-temporal data, and can model and analyze spatio-temporal data through deep learning and graph neural network technologies, and extract and output the corresponding user feature vector.

[0031] Specifically, the spatio-temporal data of the second candidate user group can have both time and space dimensions at the same time. For example, the spatial data with time elements and changing with time. The geographical locations of users of different ages in the second candidate user group change with time. All the time and space data are aggregated into spatio-temporal data and input into the intelligent large model. The intelligent large model processes and analyzes the spatio-temporal data to obtain the user feature vector of the second candidate user group. This user feature vector is used to process the spatio-temporal data of the second candidate user group into a format convenient for the mother and baby user recognition model to recognize, and includes the dimensions of time and space.

[0032] The confidence level includes that after the user feature vector is input into the mother and baby user recognition model, the mother and baby user recognition model analyzes and processes the user feature vector to obtain the probability that the first user in the second candidate user group corresponding to the user feature vector is a mother and baby user. Taking this probability as the true degree that the first user in the second candidate user group is a mother and baby user is the confidence level.

[0033] According to inputting the user feature vector into the mother and baby user recognition model to obtain the confidence level that the first user corresponding to the user feature vector is a mother and baby user, the first user with a confidence level greater than the preset threshold is used as the second customer group.

[0034] Among them, the second customer group also needs to meet two determination conditions, namely the third condition and the fourth condition. The user feature vector of the second candidate user group is recognized by the mother and baby user recognition model and the corresponding confidence level is output. Taking this confidence level greater than the preset confidence level as the third condition; and obtaining the user age of the users in the second candidate user group, and taking the user age within the preset age range as the fourth condition. The second customer group is denoted as The third condition is denoted as The fourth condition is denoted as , the second customer group can be expressed as Formula 3 as follows: Formula 3 The rule set is { , }, and the function is the second candidate user group for suspected mother and baby users to which the rule set is applied. That is, the rule set includes the third condition and the fourth condition. is the second candidate user group, and is the users who meet both the third condition and the fourth condition screened from the second candidate user group as the second customer group.

[0035] S108. Determine the first customer group and the second customer group as the mother and baby user customer group.

[0036] The first customer group is obtained from S102 , and the second customer group is obtained from S106 . Then, the final mother and baby user customer group is denoted as , which can be expressed as Formula 4, including the set of the first customer group and the second customer group. denotes the union, as follows: = Formula 4 Adopt the technical solution of the embodiment of the present application, obtain the first candidate user group of maternal and infant users and the information data of the first candidate user group. The information data includes Internet access log data and user basic data, and based on the information data, determine the first customer group; obtain the second candidate user group except the first customer group in the first candidate user group; input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the maternal and infant user recognition model to obtain the second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the maternal and infant user recognition model is used to determine the confidence level that the first user corresponding to the user feature vector is a maternal and infant user; determine the first customer group and the second customer group as the maternal and infant user customer group. It can be seen that through the information data of the first candidate user group of maternal and infant users, the first customer group that is a maternal and infant user is screened out, and then spatio-temporal data is obtained from the remaining users of the first candidate user group and input into the intelligent large model to obtain the user feature vector. By inputting the user feature vector into the maternal and infant user recognition model, the second customer group that is a maternal and infant user among the remaining users of the first candidate user group is screened out. Through the two screening methods, the first customer group and the second customer group are determined as the maternal and infant user customer group, which can solve the problem of only considering two dimensions of user Internet access logs and user basic information, make full use of multi-dimensional data and capabilities such as intelligent large models and maternal and infant user recognition models, and solve the problem of low accuracy in identifying the maternal and infant user customer group.

[0037] In one embodiment, to obtain the first candidate user group of maternal and infant users (i.e., S102), the following steps A1 - A3 can be executed: Step A1, obtain the base station location connected by the second user within a preset time period, and determine the location data of the second user.

[0038] The preset time period can be a specific time period, for example, within the past 1 - 3 years.

[0039] The second user includes: all users obtained from big data.

[0040] The base station location connected by the second user refers to the location of the base station connected by the second user, as well as the azimuth angle and signal strength of the second user relative to the base station. The operator locates the second user through the base station location connected by the second user to obtain the location information of the second user. The location information includes the location of the second user's location and the corresponding time, and the location information is determined as the location data of the second user.

[0041] That is, when the second user walks and moves, the accurate location of the second user changing with time can also be located at any time according to the base station location connected by the second user, and then the location data of the second user can be obtained.

[0042] Step A2: Based on the location data, determine the time periods and frequencies of the second user's visits to the hospital during the prenatal stage and the production stage.

[0043] Due to the specific time and location patterns during the prenatal stage and the production stage of the second user, the location data obtained through Step A1 can be used to obtain the time periods and frequencies of the second user's visits to the hospital during the prenatal stage and the production stage.

[0044] Generally, during the prenatal examination stage, the time periods and frequencies of visits to the hospital are specifically manifested as more than 5 visits to maternal and child-related hospitals in 3 consecutive months; during the production stage, the time periods and frequencies of visits to the hospital are specifically manifested as staying in the hospital continuously for 2 to 4 days within 9 months after the date of the last prenatal examination behavior, and staying at least 8 hours per day.

[0045] Step A3: According to the time periods and frequencies, determine the first candidate user group of mother and baby users from the second users.

[0046] According to the time periods and frequencies of the second user's visits to the hospital during the prenatal stage and the production stage obtained in Step A2, determine the second users who are equal to or greater than the preset time periods and preset frequencies as the first candidate user group of mother and baby users based on the user's location data.

[0047] As an example, obtain the location information of the second user in the past 1 - 3 years. According to the location information, determine the time periods and frequencies of the second user's visits to the hospital. During the prenatal stage, the preset time period for visiting relevant hospitals can be 3 consecutive months, and the preset frequency can be 5 times. During the production stage, the preset time period for visiting relevant hospitals can be within 9 months after the date of the last prenatal examination behavior, staying in the hospital continuously for 2 to 4 days, and staying at least 8 hours per day. The preset frequency includes at least once. According to the time periods and frequencies, determine the second users who meet the conditions as the first candidate user group of mother and baby users from the second users.

[0048] Specifically, the extraction rule of the first candidate user group simulates the prenatal examination and production processes of mother and baby users. Since mother and baby users generally refer to the parents of infants aged 1 - 3 years old, the location data of the second user in the past 1 - 3 years is extracted, and combined with the time and location information in the location data, the second candidate user group is selected. As follows, the processes of determining the first candidate user group of mother and baby users during the prenatal stage and the production stage are introduced separately. During the prenatal stage and the production stage, in the implementation process of screening the second user, according to the area of interest (AOI) delimited by maternal and child-related hospitals, determine the second user's visit to the hospital and calculate the length of stay of the second user in the hospital. The time point when the second user can enter the hospital AOI based on the obtained location data of the second user , determine that the second user arrives at the hospital at this time point, and the departure time point is , the length of stay of the second user in the hospital is .

[0049] Prenatal stage: Taking the current time node as the reference, within the time period from the same period three years ago to the same period one year ago, relying on the location data of the second user, screen the second users who have visited maternal and child-related hospitals more than 5 times in 3 consecutive months as the user group that meets the location characteristics of prenatal examinations. At the same time, to avoid extracting other users not related to prenatal examinations, limit the length of stay of the second user in the hospital each time to between 1 and 5 hours, and limit that the prenatal examination behaviors that occur before and after are in the same hospital. In addition, the date of the last prenatal examination behavior of the second user in the above 3 consecutive months is d, for inferring subsequent production behaviors.

[0050] Production stage: Based on a part of the second user group determined in the prenatal stage, screen the users who have continuously stayed in the hospital for 2 to 4 days within 9 months after the date of the last prenatal examination behavior, and stay at least 8 hours a day, as a small part of the second user group that meets the production characteristics. Determine this small part of the second user group as the first candidate user group of maternal and child users, denoted as S, and the second user is denoted as u, the relationship For the extraction rule of the first candidate user group, the extraction rule is the time period and number of times that the second user meets to go to the hospital in the prenatal stage and the production stage. The first candidate user group is expressed as Formula 5, as follows: Formula 5 In this embodiment, by obtaining the base station location of the second user within the preset time period, determining the location data of the second user according to the signal strength and azimuth angle of the second user relative to the base station, and determining the time period and number of times that the second user goes to the hospital in the prenatal stage and the production stage according to the location data, thereby determining the first candidate user group of maternal and child users among the second users. It can improve the accuracy of positioning the first candidate user group of maternal and child users.

[0051] In one embodiment, based on the information data, determine the first customer group (i.e., S102), and the following steps B1 - B3 can be specifically executed: Step B1, determine the user age in the first candidate user group according to the user basic data in the information data of the first candidate user group.

[0052] Since the information data of the first candidate user group includes the user basic data in the first candidate user group, according to the user basic data, basic information such as the user age, residence location, and geographical location of each user in the first candidate user group can be determined. Therefore, the user age in the first candidate user group can be determined according to the user basic data.

[0053] Step B2: Determine the number of times the users in the first candidate user group access the maternal and infant software based on the Internet access log data in the information data of the first candidate user group.

[0054] Since the information data of the first candidate user group also includes the Internet access log data of the users in the first candidate user group, based on the Internet access log data, it is possible to obtain the Internet access information of the users on maternal and infant-related software, websites, mini-programs, etc. within a preset time period. The Internet access information includes, for example, the frequency and time of browsing maternal and infant-related software. Therefore, the number of times the users in the first candidate user group access the maternal and infant software can be determined based on the Internet access log data.

[0055] Step B3: Determine the first customer group based on the user age and the number of times the users access the maternal and infant software.

[0056] Within a preset time period, obtain the user age and the number of times the users in the first candidate user group access the maternal and infant software according to Step B1 and Step B2. The preset time period can be the same as the preset time period in Step A1. Take the number of times the users access the maternal and infant software being greater than a preset quantity as the first condition, and the user age being within a preset age range as the second condition. According to Formula 1 in S102, determine the users in the first candidate user group who simultaneously meet the first condition and the second condition as the first customer group.

[0057] In this embodiment, when screening the first customer group from the first candidate user group, according to the information data of the first candidate user group, obtain the user age and the number of times the users in the first candidate user group access the maternal and infant software. When the user age and the number of times the users access the maternal and infant software meet the preset conditions, determine the set of users in the first candidate user group that meet the preset conditions as the first customer group of maternal and infant users, which can accurately obtain the first customer group in the first candidate user group.

[0058] In one embodiment, input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vectors of the second candidate user group (i.e., S106), and the following steps C1 - C2 can be executed: Step C1: Obtain the spatio-temporal data of the second candidate user group and input the spatio-temporal data into the intelligent large model.

[0059] Among them, the spatio-temporal data includes: the information data and location data of the second candidate user group in terms of time and space, and the intelligent large model includes a spatio-temporal large model.

[0060] Specifically, since only a small number of mother and baby users generate relevant mobile Internet logs during the parenting period, identifying mother and baby users based on user Internet log data will result in a low coverage rate. Therefore, it is necessary to obtain the remaining users outside the first customer group from the first candidate user group as the second candidate user group. By obtaining the spatio-temporal data of each user in the second candidate user group, the information data and location data of the second candidate user group in terms of time and space are determined. For example, obtain the base station locations connected by the target user in the second candidate user group during a preset time period, and then determine location data such as the number of times the target user goes to the hospital during the prenatal stage and the production stage, and obtain the information data of the target user. Determine the age of the target user based on the basic user data in the information data, and determine the number of times the user accesses the mother and baby software based on the Internet log data. It should be noted that the spatio-temporal data may also include other data of the users in the second candidate user group in terms of time and space, and no specific limitation is made. During the preset time period, as time changes, the information data and location data of the users in the second candidate user group also change accordingly. At different times and in different spaces, the user information in the second candidate user group is constantly changing dynamically. The spatio-temporal data not only includes the geographical location information of the users, but also records the spatial data that changes over time.

[0061] The intelligent large model can be a spatio-temporal large model, and the spatio-temporal large model is an artificial intelligence pre-training model constructed based on the ultra-large-scale dynamic graph neural network technology.

[0062] Step C2, convert the spatio-temporal data into user feature vectors of the second candidate user group through the intelligent large model.

[0063] Input the spatio-temporal data into the intelligent large model to generate a format that is convenient for the mother and baby user identification model to identify in both the time and space dimensions, that is, output user feature vectors through the intelligent large model.

[0064] The intelligent large model includes a spatio-temporal large model. For example, the Jiutian Chuanliu Travel Large Model, that is, the Chuanliu Large Model. As an example, the Chuanliu Large Model can collect the spatio-temporal data of all users in the second candidate user group from the same period 3 years ago to the same period 1 year ago at the current time point, and the time interval between the front and back of each user's spatio-temporal tuple is 10 days. Submit the collected user spatio-temporal stream to the open interface of the Chuanliu Large Model, and output the user feature vector of each user, where the dimension of the user feature vector is 768.

[0065] In this embodiment, by obtaining the spatio-temporal data of the second candidate user group, the change data of the second candidate user group in terms of time and space is obtained. The spatio-temporal data is input into the spatio-temporal large model, and the spatio-temporal large model converts the spatio-temporal data into user feature vectors corresponding to the users in the second candidate user group. In this way, the changes in time and space of the users in the second candidate user group can be displayed through the user feature vectors, which is beneficial for subsequent training of downstream tasks.

[0066] In one embodiment, the user feature vectors are input into the mother and baby user recognition model to obtain the second customer group (i.e., S106), and the following steps D1 - D2 can be executed: Step D1, input the user feature vectors into the mother and baby user recognition model to obtain the output result of the user feature vectors.

[0067] Among them, the output result includes: the confidence that the first user corresponding to the user feature vector is a mother and baby user.

[0068] The first user is the user corresponding to the user feature vector output by the intelligent large model in step C2.

[0069] The mother and baby user recognition model includes a trained discriminant model for determining the confidence that the first user corresponding to the user feature vector is a mother and baby user.

[0070] Input the user feature vectors into the mother and baby user recognition model. The mother and baby user recognition model can output the probability that the first user corresponding to the user feature vector is a mother and baby user, and this output result, i.e., the probability, is used as the confidence for judging that the first user is a mother and baby user.

[0071] Step D2, according to the output result, obtain the first users whose confidence is greater than the preset threshold, and determine the first users whose confidence is greater than the preset threshold as the second customer group.

[0072] The preset threshold includes values with relatively high confidence set according to experience, such as 0.5, 0.8, etc., and is not specifically limited.

[0073] The set of the first users whose confidence in the output result is higher than the preset threshold is used as the second customer group. The second customer group includes the mother and baby users in the second candidate user group. As shown in Formula 3, the second customer group is obtained through calculation.

[0074] In this embodiment, by inputting the user feature vector into the mother and baby user recognition model, the confidence level that the first user corresponding to the user feature vector is a mother and baby user is output, and it is determined whether the first user is a mother and baby user. It can perform a secondary screening of the mother and baby users in the first candidate user group. In addition to screening out the first customer group including mother and baby users through the information data of the first candidate user group for the first time, it can also screen the remaining second candidate user group to obtain the second customer group including mother and baby users, and more accurately determine the mother and baby user customer group.

[0075] In one embodiment, the training of the mother and baby user recognition model can execute the following steps E1 - E4: Step E1, obtain a plurality of sample training data and corresponding sample confidence levels.

[0076] The sample training data includes: positive sample data and negative sample data of mother and baby users.

[0077] The positive sample data of mother and baby users includes: using the user data obtained as the data of mother and baby users as the positive sample data, and the sample confidence level corresponding to the positive sample is relatively high. The positive sample is the sample corresponding to the sample data that is a mother and baby user.

[0078] The negative sample data of mother and baby users includes: using the user data obtained that is not the data of mother and baby users as the negative sample data, and the sample confidence level corresponding to the negative sample is relatively low. The negative sample is the sample corresponding to the sample data that is not a mother and baby user.

[0079] As an example, in the previous SMS reach service, the positive sample data of mother and baby users in the sample training data is collected. For example, the relevant data of the third user who sends notification SMS by various mother and baby product stores, nursing centers, early education institutions, etc. is collected as the positive sample data, where the third user is the user corresponding to the sample training data. Since the third users are all registered real mother and baby users, the authenticity and effectiveness of the sample training data can be guaranteed. In addition, in the process of machine learning modeling, the sample training data should also include more negative samples than positive samples. In the full user group including the third user, after removing the aforementioned first customer group and positive samples, then randomly sample three times the number of negative samples of the positive samples.

[0080] Finally, encrypt the unique identifier of the third user in the sample training data to prevent information leakage. It can also be divided into a training set and a test set according to a scale ratio of 8:2. The training set is used to train the subsequent discriminant model, and the test set is used to test the accuracy of the trained model. The confusion matrix of the number of samples is shown in Table 1. The number of positive samples in the training set is 4811, and the number of negative samples is 14061; the number of positive samples in the test set is 1222, and the number of negative samples is 3609:

[0081] Table 1: Confusion Matrix of Sample Numbers Step E2: Input the sample training data into the discriminant model set up based on the Light Gradient Boosting Machine to output the sample output result, where the sample output result includes the confidence that the third user corresponding to the sample training data is a mother and baby user.

[0082] For the modeling requirements of the problem scenario of mother and baby user identification, mapping the user feature vector to the recognition confidence can select many representative traditional machine learning models, including logistic regression, linear discriminant, classification and regression tree, gradient boosting tree and other models. Among them, since the gradient boosting tree has good adaptability to large model feature vectors, therefore, an efficient engineering implementation version of the gradient boosting tree is selected, such as selecting the Light Gradient Boosting Machine (LGBM) as the discriminant model. On the basis of using the discriminant model, training is carried out based on the collected sample training data. After the training process converges, a mother and baby user identification model is obtained.

[0083] Among them, the discriminant model based on the Light Gradient Boosting Machine includes preprocessing the obtained sample training data, such as feature encoding, missing value processing, etc., and creating a data set according to the preprocessed data.

[0084] Continuing with the above example, the hyperparameters and other settings of the training process of the discriminant model can be as shown in Table 2:

[0085] Table 2: Specific Training Settings Based on the parameters of the above-trained discriminant model, the final sample output result of the discriminant model is obtained.

[0086] Step E3: Train the discriminant model based on the sample output result and the sample confidence to obtain the trained discriminant model.

[0087] According to the sample output result obtained in Step E2, compare it with the sample confidence. By adjusting the discriminant model parameters, make the sample output result and the sample confidence not very different or consistent, and the training process of the discriminant model converges to obtain the trained discriminant model.

[0088] In addition, in order to avoid the discriminant model from overfitting in the later stage of training and missing the maximum performance point in the parameter space, an early stopping strategy is added to the discriminant model training, so that the discriminant model training stops at the maximum point in the parameter space to avoid overfitting.

[0089] Continuing with the above example, the specific settings of the early stopping strategy are shown in Table 3:

[0090] Table 3: Early stopping strategy settings Among them, the evaluation metric can be a curve for evaluating the discriminative model during training, used to determine the maximum point of the performance of the discriminative model during training. The number of stopping rounds refers to the number of times the discriminative model stops training. The evaluation round interval refers to the interval for evaluating the discriminative model and training the sample training data of the discriminative model, that is, evaluating once for every 200 sample training data. The validation set size refers to the number of times of using the test set to validate the discriminative model.

[0091] Step E4, determine the trained discriminative model as the maternal and child user recognition model.

[0092] It should be noted that, as Figure 2 shown, it is a schematic diagram of the River Flow large model using the maternal and child user recognition model to identify maternal and child users. First, 1. Obtain the spatio-temporal data of the user according to the spatio-temporal flow of the user; 2. Input the spatio-temporal data into the River Flow large model; 3. Use the River Flow large model to convert the spatio-temporal data into a 768-dimensional user feature vector; 4. Input the user feature vector into the maternal and child user recognition model and train the user feature vector according to task labels, etc.; 5. Identify maternal and child users through the maternal and child user recognition model.

[0093] In this embodiment, the obtained sample training data includes positive sample data and negative sample data, and the disclosed process of obtaining sample training data can improve the quality of the sample training data for training, ensure the performance of the discriminative model to be trained and the effectiveness of the obtained sample training data. By inputting the sample training data into the discriminative model set based on the lightweight gradient boosting machine, obtaining the sample output result, and adjusting the parameters of the discriminative model, the discriminative model converges to obtain the trained discriminative model, which is used as the maternal and child user recognition model. Through the maternal and child user recognition model, the downstream tasks corresponding to the intelligent large model can be executed, improving the accuracy of identifying the maternal and child user customer group.

[0094] Figure 3 is a schematic flowchart of a user recognition method according to another embodiment of the present application. As Figure 3 shown, the method includes the following steps: S301, obtain the base station locations connected by the user within a preset time period, and determine the location data of the user.

[0095] S302, based on the location data, determine the time periods and frequencies of the user going to the hospital during the prenatal stage and the production stage; according to the time periods and frequencies, determine the first candidate user group of maternal and child users.

[0096] S303, obtain the information data of the first candidate user group, and the information data includes Internet access log data and user basic data.

[0097] S304. Based on the information data, determine the user's age and the number of times the user accesses the maternal and child software, and determine the users in the first candidate user group that meet the preset conditions as the first customer group.

[0098] The preset conditions may include: taking the number of times the user accesses the maternal and child software being greater than the preset quantity as the first condition, and taking the user's age being within the preset age range as the second condition. The set of users who meet both the first condition and the second condition is the first customer group.

[0099] S305. Obtain the second candidate user group among the first candidate user group except for the first customer group.

[0100] S306. Obtain the spatio-temporal data of the second candidate user group and input the spatio-temporal data into the intelligent large model.

[0101] S307. Through the intelligent large model, convert the spatio-temporal data into the user feature vectors of the second candidate user group.

[0102] S308. Through training the downstream task, train a discriminant model set based on the lightweight gradient boosting machine to obtain a maternal and child user recognition model.

[0103] S309. Input the user feature vectors of the second candidate user group into the maternal and child recognition model to obtain the confidence that the first user corresponding to the user feature vectors is a maternal and child user.

[0104] S310. Determine the users with a confidence greater than the preset threshold as the second customer group.

[0105] S311. Determine the set of the first customer group and the second customer group as the maternal and child user customer group.

[0106] The specific processes of the above S301 to S311 have been described in detail in the above embodiments and will not be elaborated here.

[0107] Adopt the technical solution of the embodiment of the present application to obtain the first candidate user group of maternal and child users and the information data of the first candidate user group. The information data includes Internet access log data and user basic data, and based on the information data, determine the first customer group; obtain the second candidate user group except the first customer group in the first candidate user group; input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the maternal and child user identification model to obtain the second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the maternal and child user identification model is used to determine the confidence level that the first user corresponding to the user feature vector is a maternal and child user; determine the first customer group and the second customer group as the maternal and child user customer group. It can be seen that through the information data of the first candidate user group of maternal and child users, the first customer group that is a maternal and child user is screened out, and then the spatio-temporal data is obtained from the remaining users of the first candidate user group and input into the intelligent large model to obtain the user feature vector. By inputting the user feature vector into the maternal and child user identification model, the second customer group that is a maternal and child user among the remaining users of the first candidate user group is screened out. Through the two screening methods, the first customer group and the second customer group are determined as the maternal and child user customer group, which can solve the problem of only considering two dimensions of user Internet access logs and user basic information, make full use of multi-dimensional data and capabilities such as intelligent large models and maternal and child user identification models, and solve the problem of low accuracy in identifying maternal and child user customer groups.

[0108] In summary, specific embodiments of the present subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the acts recited in the claims can be performed in a different order and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing may be advantageous.

[0109] The above is a method for user identification provided by an embodiment of the present application. Based on the same idea, an embodiment of the present application also provides a device for user identification.

[0110] Figure 4 It is a schematic structural diagram of a device for user identification according to an embodiment of the present invention. As Figure 4 shown, the device for user identification includes: a first acquisition module 41, a second acquisition module 42, an identification module 43, and a determination module 44: The first acquisition module 41 is configured to acquire the first candidate user group of maternal and child users and the information data of the first candidate user group. The information data includes Internet access log data and user basic data, and based on the information data, determine the first customer group; The second acquisition module 42 is configured to acquire a second candidate user group excluding the first customer group from the first candidate user group; The recognition module 43 is configured to input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the mother and baby user recognition model to obtain the second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the mother and baby user recognition model is used to determine the confidence level that the first user corresponding to the user feature vector is a mother and baby user; The determination module 44 is configured to determine the first customer group and the second customer group as the mother and baby user customer group.

[0111] In one embodiment, the first acquisition module 41 is specifically configured to acquire the base station locations connected by the second user within a preset time period, and determine the location data of the second user; based on the location data, determine the time period and frequency of the second user going to the hospital during the prenatal stage and the production stage; according to the time period and frequency, determine the first candidate user group of mother and baby users from the second users.

[0112] In one embodiment, the first acquisition module 41 is specifically further configured to determine the user ages in the first candidate user group according to the user basic data in the information data of the first candidate user group; determine the number of times the users in the first candidate user group access the mother and baby software according to the Internet access log data in the information data of the first candidate user group; based on the user ages and the number of times the users access the mother and baby software, determine the first customer group.

[0113] In one embodiment, the recognition module 43 is specifically configured to acquire the spatio-temporal data of the second candidate user group, and input the spatio-temporal data into the intelligent large model, where the spatio-temporal data includes: information data and location data of the second candidate user group in time and space, and the intelligent large model includes a spatio-temporal large model; through the intelligent large model, convert the spatio-temporal data into the user feature vector of the second candidate user group.

[0114] In one embodiment, the recognition module 43 is specifically further configured to input the user feature vector into the mother and baby user recognition model to obtain the output result of the user feature vector, and the output result includes: the confidence level that the first user corresponding to the user feature vector is a mother and baby user; according to the output result, obtain the first users with a confidence level greater than the preset threshold, and determine the first users with a confidence level greater than the preset threshold as the second customer group.

[0115] In one embodiment, the device further includes a training module, which specifically includes: An acquisition unit, configured to acquire a plurality of sample training data and corresponding sample confidence levels, where the sample training data includes: positive sample data and negative sample data of mother and baby users; An output unit, configured to input sample training data into a discrimination model set based on a lightweight gradient boosting machine, and output a sample output result, where the sample output result includes: the confidence that the third user corresponding to the sample training data is a mother and baby user; A training unit, configured to train the discrimination model based on the sample output result and the sample confidence, to obtain a trained discrimination model; A determination unit, configured to determine the trained discrimination model as a mother and baby user identification model.

[0116] By adopting the technical solution of the embodiment of the present application, a first candidate user group of mother and baby users and information data of the first candidate user group are obtained, where the information data includes Internet access log data and user basic data, and based on the information data, a first customer group is determined; a second candidate user group other than the first customer group in the first candidate user group is obtained; the spatio-temporal data of the second candidate user group is input into an intelligent large model to obtain a user feature vector of the second candidate user group, and the user feature vector is input into the mother and baby user identification model to obtain a second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the mother and baby user identification model is used to determine the confidence that the first user corresponding to the user feature vector is a mother and baby user; the first customer group and the second customer group are determined as the mother and baby user customer group. It can be seen that through the information data of the first candidate user group of mother and baby users, the first customer group that is a mother and baby user is screened out, and then spatio-temporal data is obtained from the remaining users of the first candidate user group and input into the intelligent large model to obtain a user feature vector. By inputting the user feature vector into the mother and baby user identification model, the second customer group that is a mother and baby user among the remaining users of the first candidate user group is screened out. Through the two screening methods, the first customer group and the second customer group are determined as the mother and baby user customer group, which can solve the problem of low accuracy in identifying the mother and baby user customer group by only considering two dimensions of user Internet access logs and user basic information, and make full use of multi-dimensional data and capabilities such as the intelligent large model and the mother and baby user identification model.

[0117] Those skilled in the art should understand that Figure 4 the user identification device in

[0118] can be used to implement the user identification method described above, and the detailed description thereof should be similar to the method part described above. To avoid redundancy, it will not be described in detail here. Figure 5Schematic diagram of the structure of an electronic device for implementing various embodiments of the present application. The electronic device may vary greatly due to different configurations or performances, and may include a processor 510, a communications interface 520, a memory 1130, and a communication bus 540. Among them, the processor 510, the communication interface 520, and the memory 530 complete communication with each other through the communication bus 540. The processor 510 can call a computer program stored in the memory 530 and executable on the processor 510 to perform the following steps: Obtain the first candidate user group of maternal and child users and the information data of the first candidate user group. The information data includes Internet access log data and user basic data, and based on the information data, determine the first customer group; Obtain the second candidate user group in the first candidate user group except the first customer group; Input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the maternal and child user recognition model to obtain the second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the maternal and child user recognition model is used to determine the confidence level that the first user corresponding to the user feature vector is a maternal and child user; Determine the first customer group and the second customer group as the maternal and child user customer group.

[0119] Adopt the technical solution of the embodiment of the present application to obtain the first candidate user group of mother and baby users and the information data of the first candidate user group. The information data includes Internet access log data and user basic data, and based on the information data, determine the first customer group; obtain the second candidate user group in the first candidate user group except the first customer group; input the spatio-temporal data of the second candidate user group into the intelligent large model to obtain the user feature vector of the second candidate user group, and input the user feature vector into the mother and baby user identification model to obtain the second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the mother and baby user identification model is used to determine the confidence level that the first user corresponding to the user feature vector is a mother and baby user; determine the first customer group and the second customer group as the mother and baby user customer group. It can be seen that through the information data of the first candidate user group of mother and baby users, the first customer group that is a mother and baby user is screened out, and then the spatio-temporal data is obtained from the remaining users of the first candidate user group and input into the intelligent large model to obtain the user feature vector. By inputting the user feature vector into the mother and baby user identification model, the second customer group that is a mother and baby user among the remaining users of the first candidate user group is screened out. Through the two screening methods, the first customer group and the second customer group are determined as the mother and baby user customer group, which can solve the problem of only considering two dimensions of user Internet access logs and user basic information, make full use of multi-dimensional data and capabilities such as intelligent large models and mother and baby user identification models, and solve the problem of low accuracy in identifying the mother and baby user customer group.

[0120] The specific implementation steps can refer to each step of the method embodiment of the above user identification, and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0121] It should be noted that the electronic devices in the embodiments of the present application include: servers, terminals, or other devices except terminals.

[0122] The above electronic device structure does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, the input unit may include a Graphics Processing Unit (GPU) and a microphone, and the display unit may be configured with a display panel in the form of a liquid crystal display, an organic light-emitting diode, etc. The user input unit includes at least one of a touch panel and other input devices. The touch panel is also called a touch screen. Other input devices may include, but are not limited to, a physical keyboard, function keys (such as volume control keys, power on / off keys, etc.), a trackball, a mouse, a joystick, which will not be elaborated here.

[0123] The memory can be used to store software programs and various data. The memory may mainly include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store an operating system, applications or instructions required for at least one function (such as a sound playback function, an image playback function, etc.). In addition, the memory may include volatile memory or non-volatile memory, or the memory may include both volatile and non-volatile memory. The non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct rambus random access memory (DRRAM).

[0124] The processor may include one or more processing units; optionally, the processor may integrate an application processor and a modem processor, wherein the application processor primarily handles operations related to the operating system, user interface, and application programs, and the modem processor primarily processes wireless communication signals, such as a baseband processor. It is understood that the modem processor may not be integrated into the processor.

[0125] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned user identification method embodiment are implemented and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0126] Among them, the processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disc, etc.

[0127] Another embodiment of the present application provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement each process of the above method embodiment for user identification, and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0128] It should be understood that the chip mentioned in the embodiments of the present application may also be referred to as a system-on-chip, system chip, chip system, or system-on-chip, etc.

[0129] Another embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the processor is used to run programs or instructions to implement each process of the above method embodiment for product recommendation, and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0130] It should be noted that in this article, the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article, or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article, or device including that element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in a reverse order according to the functions involved. For example, the described method may be executed in an order different from that described, and various steps may be added, omitted, or combined. Additionally, the features described with reference to certain examples may be combined in other examples.

[0131] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present application.

[0132] The embodiments of the present application have been described above in conjunction with the accompanying drawings. However, the present application is not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative rather than restrictive. Under the inspiration of the present application, those of ordinary skill in the art can also make many forms without departing from the purpose of the present application and the scope protected by the claims, and all of them belong to the protection scope of the present application.

Claims

1. A method for user identification, characterized in that, The method includes: Obtaining a first candidate user group of maternal and infant users and information data of the first candidate user group, where the information data includes Internet access log data and user basic data, and determining a first customer group based on the information data; Obtaining a second candidate user group among the first candidate user group excluding the first customer group; Inputting the spatio-temporal data of the second candidate user group into an intelligent large model to obtain a user feature vector of the second candidate user group, and inputting the user feature vector into a maternal and infant user recognition model to obtain a second customer group; the spatio-temporal data includes data of the second candidate user group in the time and space dimensions, and the maternal and infant user recognition model is used to determine the confidence that the first user corresponding to the user feature vector is a maternal and infant user; Determining the first customer group and the second customer group as the maternal and infant user customer group.

2. The method according to claim 1, characterized in that, The obtaining of the first candidate user group of maternal and infant users includes: Obtaining the base station locations connected by a second user within a preset time period, and determining the location data of the second user; Based on the location data, determining the time period and frequency of the second user going to the hospital during the prenatal stage and the production stage; According to the time period and the frequency, determining the first candidate user group of the maternal and infant users from the second users.

3. The method according to claim 1, characterized in that, The determining of the first customer group based on the information data includes: Determining the user ages in the first candidate user group according to the user basic data in the information data of the first candidate user group; Determining the number of times the users in the first candidate user group access the maternal and infant software according to the Internet access log data in the information data of the first candidate user group; Based on the user ages and the number of times the users access the maternal and infant software, determining the first customer group.

4. The method according to claim 1, characterized in that, The inputting of the spatio-temporal data of the second candidate user group into an intelligent large model to obtain a user feature vector of the second candidate user group includes: Obtaining the spatio-temporal data of the second candidate user group, and inputting the spatio-temporal data into the intelligent large model, where the spatio-temporal data includes the information data and location data of the second candidate user group in the time and space, and the intelligent large model includes a spatio-temporal large model; Converting the spatio-temporal data into a user feature vector of the second candidate user group through the intelligent large model.

5. The method according to claim 1, characterized in that, The inputting of the user feature vector into a maternal and infant user recognition model to obtain a second customer group includes: Inputting the user feature vector into the maternal and infant user recognition model to obtain an output result of the user feature vector, where the output result includes the confidence that the first user corresponding to the user feature vector is a maternal and infant user; According to the output result, obtaining the first users with the confidence greater than a preset threshold, and determining the first users with the confidence greater than the preset threshold as the second customer group.

6. The method according to claim 1, wherein The training of the maternal and infant user recognition model includes: Obtaining a plurality of sample training data and corresponding sample confidences, where the sample training data includes positive sample data and negative sample data of the maternal and infant users; Input the sample training data into a discriminant model set based on a lightweight gradient boosting machine, and output the sample output result, where the sample output result includes: the confidence that the third user corresponding to the sample training data is the maternal and child user; Train the discriminant model based on the sample output result and the sample confidence to obtain a trained discriminant model; Determine the trained discriminant model as the maternal and child user identification model.

7. A user identification device, characterized in that, It includes: A first acquisition module, configured to acquire a first candidate user group of maternal and child users and information data of the first candidate user group, where the information data includes Internet access log data and user basic data, and determine a first customer group based on the information data; A second acquisition module, configured to acquire a second candidate user group other than the first customer group in the first candidate user group; An identification module, configured to input the spatio-temporal data of the second candidate user group into an intelligent large model to obtain a user feature vector of the second candidate user group, and input the user feature vector into a maternal and child user identification model to obtain a second customer group; the spatio-temporal data includes: data of the second candidate user group in the time and space dimensions, and the maternal and child user identification model is used to determine the confidence that the first user corresponding to the user feature vector is the maternal and child user; A determination module, configured to determine the first customer group and the second customer group as the maternal and child user customer group.

8. An electronic device, characterized in that, It includes a processor and a memory electrically connected to the processor, where the memory stores a computer program, and the processor is configured to call and execute the computer program from the memory to implement a user identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The storage medium is used to store a computer program, and the computer program can be executed by a processor to implement a user identification method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, It includes a computer program, and when the computer program is executed by a processor, it implements a user identification method according to any one of claims 1 to 6.

Images