Customer classification method and device

Abstract

The invention relates to the field of big data, and provides a customer classification method and device, and the method comprises the steps: obtaining the customer data of a plurality of customers in a preset time period; grouping the plurality of clients according to the client attribute information; according to the existing product information and the transaction information of each customer and the behavior information of the user reaching the product, multi-dimensional evaluation parameters of each customer are calculated, and the multi-dimensional evaluation parameters comprise consumption time interval information, consumption frequency information, consumption amount information, asset information and reaching information of each product; for each group, sorting the clients in each group according to the evaluation parameters of each dimension; according to a customer sorting result of each dimension of evaluation parameter of each group, calculating scores of customers in each group in each dimension of evaluation parameter; and classifying the clients in each group according to the scores of the clients in each dimension of evaluation parameters in each group. The accuracy of customer classification can be improved.

Description

Technical Field

[0001] This article relates to the field of big data, and in particular to a customer classification method and device. Background Art

[0002] In the prior art, data point embedding is usually used to record customer behavior (including purchase records, browsing records, etc.), and customer portraits are roughly calculated based on customer behavior. The account manager analyzes the customer type based on the customer portrait and provides services to the customer based on the analysis results. This customer classification method only considers user behavior and is subjectively classified by the account manager. There are low customer classification efficiency and poor accuracy, which leads to the problem of mismatch between the services provided to users and customer needs, and the problem of failure to meet the diverse needs of customers. Summary of the invention

[0003] This paper aims to solve the problems of low efficiency and poor accuracy in customer classification in the existing technology.

[0004] In order to solve the above technical problems, the first aspect of this article provides a customer classification method, including:

[0005] Acquire customer data of multiple customers within a predetermined time period, wherein the customer data includes customer attribute information, existing product information, transaction information, and user behavior information about product access within the predetermined time period;

[0006] Grouping the multiple customers according to the customer attribute information;

[0007] Calculate the multi-dimensional evaluation parameters of each customer based on the existing product information, transaction information and user product contact behavior information of each customer, wherein the multi-dimensional evaluation parameters include: consumption time interval information, consumption frequency information, consumption amount information, asset information and product contact information;

[0008] For each group, the customers in each group are sorted according to each dimension of evaluation parameters; based on the customer sorting results of each dimension of evaluation parameters of each group, the scores of the customers in each group in each dimension of evaluation parameters are calculated;

[0009] The customers in each group are classified according to their scores in each dimension of evaluation parameters.

[0010] In a further embodiment of the present invention, the customer classification method further includes:

[0011] Processing missing values ​​and outliers on the multiple customer data to filter out customer data that does not meet the requirements;

[0012] Standardize the variables in the processed customer data;

[0013] The categorized variables in the processed customer data are coded as dummy variables.

[0014] In a further embodiment of the present invention, standardizing the variables in the processed customer data includes:

[0015] For each variable, determine the maximum and minimum values ​​of the variable in all customer data after processing;

[0016] Divide the value of the variable in each customer data after processing by the difference between the maximum and minimum values ​​to obtain the standard value of the variable.

[0017] In a further embodiment of the present invention, according to the customer ranking result of each dimension evaluation parameter of each group, the score of each customer in each group in each dimension evaluation parameter is calculated, including:

[0018] Sort the customers in each group according to the order of each dimension evaluation parameter from small to large;

[0019] According to the customer ranking results of each dimension of the evaluation parameter of each group, calculate the proportion of the ranking position of the customers in each group in the evaluation parameter ranking results to the total number of customers in the group;

[0020] The proportion of the ranking positions of customers in each group in the evaluation parameter ranking results to the total number of customers in the group is matched with the preset quantile scoring library of the evaluation parameters to determine the scores of customers in each group in each dimension of the evaluation parameters, wherein the preset quantile scoring library of each evaluation parameter records the correlation between the quantile range and the score.

[0021] In a further embodiment of the present invention, the customers in each group are classified according to the scores of the customers in each group in each dimension of the evaluation parameter, including:

[0022] Match the scores of the evaluation parameters of each dimension of the customers in each group with the customer classification library, and use the matched type as the type of the customers in each group;

[0023] The customer classification library stores the association between customer types and evaluation parameter score ranges.

[0024] In a further embodiment of the present invention, the customer classification method further includes:

[0025] According to the customer type, the information push strategy is determined by querying the strategy library, wherein the strategy library includes the association between the customer type and the strategy;

[0026] Push information to customers according to the information push strategy.

[0027] In a further embodiment of the present invention, the customer classification method further includes:

[0028] Screen out customers of the intended customer type;

[0029] Input the multi-dimensional evaluation parameters of the screened customers and the product information to be pushed into the pre-established product preference prediction model to obtain the customer's preference for the pushed product information;

[0030] Push product information to customers based on their preference for the pushed product information;

[0031] The product preference prediction model is trained based on a training sample set established based on historical response information of multiple customers of a predetermined customer type to push different products, wherein the training sample set includes the customer's multi-dimensional evaluation parameters, product information, and the customer's preference probability for the product;

[0032] The product preference prediction model includes: a support vector machine module based on a hybrid kernel function and an rank-preserving regression module.

[0033] A second aspect of the present invention provides a customer classification device, comprising:

[0034] an acquisition unit, used to acquire customer data of a plurality of customers within a predetermined time period;

[0035] A calculation unit is used to calculate the multi-dimensional evaluation parameters of each customer based on the acquired customer data, wherein the customer data includes customer attribute information, asset information, existing product information, transaction information and user product contact behavior information within a predetermined time period, and the multi-dimensional evaluation parameters include: consumption time interval information, consumption frequency information, consumption amount information, asset information and product contact information;

[0036] A scoring unit, used to calculate the customer's score in each dimension of evaluation parameters according to the customer ranking results according to each dimension of evaluation parameters;

[0037] The classification unit is used to classify customers according to their scores in each dimension of evaluation parameters.

[0038] A third aspect of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any one of the aforementioned embodiments when executing the computer program.

[0039] A fourth aspect of the present invention provides a computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor of a computer device, implements the method described in any of the foregoing embodiments.

[0040] The customer classification method and device provided in this article obtain customer data including customer attribute information, existing product information within a predetermined time period, transaction information, and user product contact behavior information; calculate multi-dimensional evaluation parameters including consumption time interval information, consumption frequency information, consumption amount information, asset information, and product contact information according to customer data; pre-group customers according to customer attribute information, and for each group, sort the customers in each group according to each dimensional evaluation parameter; calculate the score of each customer in each group in each dimensional evaluation parameter according to the customer sorting result of each dimensional evaluation parameter of each group; classify customers in each group according to the score of each customer in each dimensional evaluation parameter of each group, which can comprehensively consider various aspects of customer information, first group customers using relatively fixed customer attribute information, and then further classify each group using relatively personalized customer evaluation parameters, which can eliminate the classification error caused by classifying customers with different attributes together, so that the relative classification result of customers in the predetermined time period is more accurate. This provides a basis for accurate product push and targeted services to customers.

[0041] In order to make the above and other purposes, features and advantages of this article more obvious and easy to understand, the following specifically cites preferred embodiments and describes them in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of this article or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of this article. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A flow chart showing a method for classifying customers in an embodiment of this invention is shown;

[0044] Figure 2 Another flow chart of the customer classification method of the embodiment of this article is shown;

[0045] Figure 3 A flowchart showing the calculation process of equally dividing the evaluation parameters of each dimension for customers in each group in the embodiment of this article is shown;

[0046] Figure 4 A flow chart showing the information push process of the embodiment of this article is shown;

[0047] Figure 5 A flowchart showing a product preference probability prediction process in an embodiment of this invention is shown;

[0048] Figure 6 A structural diagram of a customer classification device according to an embodiment of this invention is shown;

[0049] Figure 7 A structural diagram of a computer device according to an embodiment of this invention is shown.

[0050] Description of the accompanying symbols:

[0051] 601, acquisition unit;

[0052] 602, computing unit;

[0053] 603, scoring unit;

[0054] 604, taxonomic unit;

[0055] 702. Computer equipment;

[0056] 704, processor;

[0057] 706. Memory;

[0058] 708, driving mechanism;

[0059] 710, input / output module;

[0060] 712. Input devices;

[0061] 714. Output devices;

[0062] 716. Presentation equipment;

[0063] 718. Graphical user interface;

[0064] 720, network interface;

[0065] 722. Communication link;

[0066] 724. Communication bus. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of this article to clearly and completely describe the technical solutions in the embodiments of this article. Obviously, the described embodiments are only part of the embodiments of this article, not all of the embodiments. Based on the embodiments of this article, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this article.

[0068] It should be noted that the terms "first", "second", etc. in the specification and claims of this article and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of this article described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or equipment that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0069] This specification provides method operation steps as described in the embodiments or flow charts, but more or fewer operation steps may be included based on routine or non-creative work. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the system or device product is executed in practice, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel.

[0070] It should be noted that the customer classification method and device in this article can be used in the financial field, and can also be used in any field other than the financial field. The application field of the customer classification method and device in this article is not limited.

[0071] In one embodiment of the present invention, a customer classification method is provided to solve the problem of low efficiency and poor accuracy of customer classification in the prior art. Figure 1 As shown, customer classification methods include:

[0072] Step 101, obtaining customer data of multiple customers within a predetermined time period, wherein the customer data includes customer attribute information, existing product information, transaction information, and user behavior information regarding product access within the predetermined time period.

[0073] Specifically, the customer attribute information includes age, education level, occupation, and customer registration time (years). The customer registration time refers to the customer registration time in the institution that provides the customer classification method, such as the customer registration time in a banking institution.

[0074] The predetermined time period may be, for example, one year, one quarter or one month, and may be set according to actual conditions, and is not limited in this document.

[0075] The existing product information is related to the application field of the customer classification method of this case. For example, when this case is applied to the field of financial technology, the existing product information is the balance of the customer's wage payment product and the amount of various financial products. Among them, various financial products include but are not limited to personal consumption loans, personal housing loans, deposits, treasury bonds, funds, bonds, RMB and foreign currency wealth management, securities depository accounts, credit card overdrafts, credit card consumption, current deposits, time deposits, account foreign exchange, whether online banking is opened, whether Lingtong card is opened, whether provident fund is opened, whether mobile banking is opened, etc.

[0076] Transaction information includes but is not limited to: counter transaction information, online banking transaction information, mobile banking login and transaction information, inter-bank different-name transfer transaction information, inter-bank same-name transfer transaction information, etc. All types of transaction information include transaction time information.

[0077] Product browsing information includes: product feature information, product browsing time information, etc.

[0078] Step 102: group multiple customers according to customer attribute information.

[0079] When this step is implemented, each parameter in the customer attribute information may be divided into ranges, and each range of all parameters in the customer attribute information may be combined in any way to obtain customer groups.

[0080] Step 103, calculate the multi-dimensional evaluation parameters of each customer based on the existing product information, transaction information and user product access behavior information of each customer, wherein the multi-dimensional evaluation parameters include: consumption time interval information (Recency, referred to as R), consumption frequency information (Frequency, referred to as F), consumption amount information (Monetary, referred to as M), asset information (Property, referred to as P) and each product access information (Browsability, referred to as B).

[0081] Among them, the consumption time interval information calculation process includes: calculating the time interval between two adjacent transactions based on the customer's transaction information within the predetermined time period, and the time interval between the two most recent transactions can be used as the customer's consumption time interval information, and the average time interval between every two transactions in all transactions generated within the customer's predetermined time period can also be used as the consumption time interval information.

[0082] The consumption frequency information calculation process includes: calculating the number of consumptions according to the transaction information of the customer within a predetermined time period, and using the number of consumptions as the consumption frequency information or using the value after dividing the number of consumptions by the predetermined time period as the consumption frequency information.

[0083] The consumption amount information calculation process includes: calculating the total consumption amount according to the transaction information of the customer within a predetermined time period, and using the total consumption amount as the consumption amount information.

[0084] The asset information calculation process includes: marking each customer's existing product information. Specifically, non-transferred products (such as deposits) are marked as positive, and transferred products (such as loans) are marked as negative.

[0085] Each product contact behavior includes product browsing, product purchase, product search, etc. Each product contact information includes: classifying the customer's contact behavior information of the product within the customer's scheduled time period, counting the number of contacts and contact duration of the same type of products under each category; determining the product contact information based on the number of contacts and contact duration of the product. In specific implementation, weights can be assigned to similar products according to their importance, and the number of contacts and contact duration can be weighted to obtain the contact information of similar products. When the customer's product contact behavior includes contact with multiple types of products, the contact information of each type of product is added and calculated to obtain the product contact information.

[0086] Step 104, for each group, sort the customers in each group according to each dimension of evaluation parameter; and calculate the score of the customers in each group in each dimension of evaluation parameter according to the customer sorting result of each dimension of evaluation parameter of each group.

[0087] Assuming that the number of customers in a certain group is 1000, these 1000 customers are sorted separately according to the size of each dimension of evaluation parameters required.

[0088] In order to eliminate the influence of the number of customers within a predetermined time period on each analysis result, the customer ranking result may be the proportion of the ranking of each customer in the ranking result to the total number of customers in the group.

[0089] Step 105: Classify the customers in each group according to the scores of the customers in each group in each dimension of the evaluation parameter.

[0090] This embodiment comprehensively considers various aspects of customer information (including customer attribute information, existing product information, transaction information and user product access behavior information within a predetermined time period), first uses relatively fixed customer attribute information to group customers, and then uses relatively personalized customer evaluation parameters to further classify each group, which can eliminate the classification error caused by classifying customers with different attributes together, making the relative classification results of customers within a predetermined time period more accurate. This provides a basis for accurate product push and targeted services to customers.

[0091] In one embodiment of the present invention, in order to ensure correct classification, a pre-processing process is also included before grouping customers, specifically as follows: Figure 2As shown, including:

[0092] Step 201 , processing missing values ​​and outliers on multiple customer data, and filtering out customer data that does not meet the requirements.

[0093] Specifically, common methods for dealing with missing values ​​include sample deletion, interpolation, modeling and prediction, etc. This article uses direct deletion to deal with missing values ​​in the data. Since the missing values ​​in the data set in this activity are a large number of missing sample characteristics of individual customers, and the sample size is large enough, it is more effective to simply delete a small part of the samples to achieve the goal.

[0094] Outliers, also known as isolated points, are values ​​that deviate from the normal range. The primary task of outlier processing is to detect isolated points, and then use sample deletion or the same method as missing value processing to process them. This paper adopts the Isolated Senli algorithm, so that abnormal samples can be isolated through fewer random feature segmentation times compared to ordinary samples, thereby removing outliers. In a specific implementation, this paper processes 308,727 samples through outlier processing, leaving 277,854 samples.

[0095] Step 202: standardize the variables in the processed customer data.

[0096] When this step is implemented, for each variable, determine the maximum and minimum values ​​of the variable in all customer data after processing; divide the value of the variable in each customer data after processing by the difference between the maximum and minimum values ​​to obtain the standard value of the variable.

[0097] Step 203: perform dummy variable encoding processing on the categorized variables in the processed customer data.

[0098] A dummy variable is an artificial variable, usually with a value of 0 or 1, to reflect the different attributes of a variable. For an independent variable with n categorical attributes, it is usually necessary to select one category as a reference, so n-1 dummy variables can be generated. Before modeling, this paper encodes all categorical variables into dummy variables to give them numerical attributes and improve the efficiency of subsequent data grouping.

[0099] In one embodiment of this invention, Figure 3 As shown, in the above step 104, according to the customer ranking result of each dimension evaluation parameter of each group, the score of each customer in each group in each dimension evaluation parameter is calculated, including:

[0100] Step 301, sorting the customers in each group according to the order of each dimension evaluation parameter from small to large.

[0101] Step 302, based on the customer ranking results of each dimension of the evaluation parameter of each group, calculate the proportion of the ranking positions of the customers in each group in the evaluation parameter ranking results to the total number of customers in the group.

[0102] Step 303, matching the proportion of the ranking position of the customers in each group in the evaluation parameter ranking result to the total number of customers in the group with the preset percentile scoring library of the evaluation parameters, to determine the score of the customers in each group in each dimension of the evaluation parameters.

[0103] Among them, the preset quantile scoring library of each evaluation parameter records the correlation between the quantile range and the score. The quantile range can be set according to actual needs, and this article does not limit its specific value.

[0104] In a specific implementation, the preset quantile scoring library of the evaluation parameters is shown in Table 1.

[0105] Table 1

[0106]

[0107]

[0108] In one embodiment of the present invention, the above step 105 classifies the customers in each group according to the scores of the customers in each evaluation parameter dimension in each group, including:

[0109] The scores of the evaluation parameters of each dimension of the customers in each group are matched with the customer classification library, and the matched type is used as the type of the customers in each group. The customer classification library stores the association between the customer type and the evaluation parameter score range.

[0110] In a specific implementation, the customer classification library is shown in Table 2.

[0111] Table 2

[0112] Evaluation parameter score range Customer Type R, F, M, P, B values ​​are all 1 or 2 Important to keep customers Four values ​​are 1 or 2, and one value is 3 or 4 Important development customers Three values ​​are 1 or 2, two values ​​are 3 or 4 Important customer retention Two values ​​are 1 or 2, three values ​​are 3 or 4 Generally keep customers One value is 1 or 2, four values ​​are 3 or 4 General development customers All five values ​​are 3 or 4 General customer retention

[0113] In Table 2, R, F, M, P, and B correspond to consumption time interval information, consumption frequency information, consumption amount information, asset information, and product reach information, respectively. Table 2 is only an example. In specific implementation, the correlation between the evaluation parameter score range and the customer type can be set according to business needs.

[0114] In other embodiments, the scores of each customer in each dimension of the evaluation parameter in each group may be weighted and summed to obtain the customer's total score, and the customers may be classified according to the customer's total score.

[0115] In one embodiment of this invention, Figure 4 As shown, on the basis of customer classification, it also includes:

[0116] Step 401: Determine the information push strategy by querying the strategy library according to the customer type.

[0117] The strategy library includes the relationship between customer types and strategies. The strategy library can be designed according to customer types, which is not limited in this article.

[0118] Step 402: Push information to the client according to the information push strategy.

[0119] In specific implementation, the methods of pushing information include but are not limited to sending text messages, emails, phone calls, APP push, etc.

[0120] In one embodiment of this invention, Figure 5 As shown, it also includes:

[0121] Step 501, screening out customers of a predetermined customer type.

[0122] Continuing with the example in the previous article, among the six customer groups divided in Table 2, the predetermined customer types are important retained customers, important development customers, and important retained customers.

[0123] Step 502: Input the screened multi-dimensional evaluation parameters of the customers and the product information to be pushed into a pre-established product preference prediction model to obtain the customer's preference for the product information to be pushed.

[0124] The recommended product information includes product feature information. Specifically, any feature that can distinguish different products can be used as product feature information.

[0125] Step 503: Push product information to customers based on their preference for pushing product information.

[0126] The product preference prediction model is trained based on a training sample set established based on the historical response information of multiple customers of a predetermined customer type to push different products, wherein the training sample set includes the customer's multi-dimensional evaluation parameters, the pushed product information, and the customer's preference probability for the product. The multi-dimensional evaluation parameters and the pushed product information are the model inputs, and the customer's preference probability for the product is the output.

[0127] The product preference prediction model includes: a support vector machine module (SVM) based on a hybrid kernel function and an isotonic regression module (Isotonic regression).

[0128] SVM introduces kernel functions to implicitly map the original spatial data into a high-dimensional feature space, and trains a linear classifier in it, so as to implement nonlinear algorithms without increasing the complexity of the algorithm. The kernel functions currently available for SVM can be divided into global kernel functions and local kernel functions. The global kernel function has global characteristics, allowing data points that are far apart to affect the value of the kernel function, while the local kernel function has local characteristics, only allowing data points that are very close to each other to affect the value of the kernel function, which can be represented by radial basis functions. This paper innovatively combines the global kernel function sigmoid and the local kernel function to form a hybrid kernel function, taking advantage of the advantages of the two types of kernel functions to reduce the impact of defects:

[0129]

[0130] Among them, K new (x,z) is the mixed kernel function, s is the parameter, K Sigmoid is the global kernel function, K RBF is a local kernel function (radial basis function RBF), x represents the input value (multidimensional evaluation parameters and product information), and z represents the output value (product preference degree).

[0131] When σ2 approaches 0, It approaches 1, which increases the weight value of the sigmoid kernel function and reduces the weight value of the RBF, thus improving its generalization ability.

[0132] The output of the model takes the probability of the dependent variable z being predicted to be 1 as the prediction score. Since the support vector machine itself does not have a probabilistic meaning, an Isotonic regression is trained here to minimize the following function to calibrate the probability of the model output:

[0133]

[0134] Where n is the number of samples, x i is the input feature of the i-th sample, y i is the target output of the i-th sample, w i is the sample weight.

[0135] In one embodiment of the present invention, customer feedback information on recommended products (including but not limited to browsing, query and other information) is also received, the customer's preference probability for the product is corrected based on the customer's feedback information on the product, and the customer's multi-dimensional evaluation parameters, pushed product information and the customer's preference probability for the product are added to the incremental sample library.

[0136] When the incremental sample library reaches a predetermined value, the model is retrained using the incremental samples to improve the accuracy of the model.

[0137] During specific implementation, the customer's multidimensional evaluation parameters may also be updated based on the customer data, and samples may be generated based on the updated multidimensional evaluation parameters and added to the incremental sample library.

[0138] Based on the same inventive concept, this article also provides a customer classification device, as described in the following embodiments. Since the principle of solving the problem by the customer classification device is similar to that of the customer classification method, the implementation of the customer classification device can refer to the customer classification method, and the repeated parts will not be repeated.

[0139] Specifically, Figure 6 As shown, the customer classification device comprises:

[0140] An acquisition unit 601 is used to acquire customer data of multiple customers within a predetermined time period;

[0141] The calculation unit 602 is used to calculate the multi-dimensional evaluation parameters of each customer based on the acquired customer data, wherein the customer data includes customer attribute information, asset information, existing product information, transaction information and user product contact behavior information within a predetermined time period, and the multi-dimensional evaluation parameters include: consumption time interval information, consumption frequency information, consumption amount information, asset information and product contact information;

[0142] The scoring unit 603 is used to calculate the customer's score in each dimension of evaluation parameters according to the customer ranking results according to each dimension of evaluation parameters;

[0143] The classification unit 604 is used to classify customers according to the scores of the customers in each dimension of evaluation parameters.

[0144] This embodiment can comprehensively consider all aspects of customer information, first use relatively fixed customer attribute information to group customers, and then use relatively personalized customer evaluation parameters to further classify each group, which can eliminate the classification error caused by classifying customers with different attributes together, making the relative classification results of customers within a predetermined time period more accurate. This provides a basis for accurate product push and targeted services to customers.

[0145] In one embodiment of the present invention, a computer device is also provided, such as Figure 7As shown, the computer device 702 may include one or more processors 704, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. The computer device 702 may also include any memory 706, which is used to store any kind of information such as code, settings, data, etc. Non-limitingly, for example, the memory 706 may include any one or more combinations of the following: any type of RAM, any type of ROM, flash memory device, hard disk, optical disk, etc. More generally, any memory may use any technology to store information. Further, any memory may provide volatile or non-volatile retention of information. Further, any memory may represent a fixed or removable component of the computer device 702. In one case, when the processor 704 executes an associated instruction stored in any memory or a combination of memories, the computer device 702 may perform any operation of the associated instruction. The computer device 702 also includes one or more drive mechanisms 708 for interacting with any memory, such as a hard disk drive mechanism, an optical disk drive mechanism, etc.

[0146] The computer device 702 may also include an input / output module 710 (I / O) for receiving various inputs (via input devices 712) and for providing various outputs (via output devices 714). A specific output mechanism may include a presentation device 716 and an associated graphical user interface 718 (GUI). In other embodiments, the input / output module 710 (I / O), input device 712, and output device 714 may not be included, and the computer device 702 may be used as a computer device in a network. The computer device 702 may also include one or more network interfaces 720 for exchanging data with other devices via one or more communication links 722. One or more communication buses 724 couple the components described above together.

[0147] The communication link 722 may be implemented in any manner, for example, through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication link 722 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc. governed by any protocol or combination of protocols.

[0148] The embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are executed.

[0149] The embodiments of this document also provide a computer-readable instruction, wherein when a processor executes the instruction, the program therein causes the processor to execute the method of any of the aforementioned embodiments.

[0150] It should be understood that in the various embodiments of this document, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this document.

[0151] It should also be understood that in the embodiments of this article, the term "and / or" is only a description of the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0152] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this article.

[0153] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0154] In the several embodiments provided herein, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or can be electrical, mechanical or other forms of connection.

[0155] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of this article.

[0156] In addition, each functional unit in each embodiment of this invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional unit.

[0157] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this article is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of this article. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0158] Specific embodiments are used in this article to illustrate the principles and implementation methods of this article. The description of the above embodiments is only used to help understand the methods and core ideas of this article. At the same time, for general technicians in this field, according to the ideas of this article, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on this article.

Claims

1. A customer classification method, characterized in that: include: Acquire customer data of multiple customers within a predetermined time period, wherein the customer data includes customer attribute information, existing product information, transaction information, and user behavior information about product access within the predetermined time period; Grouping the multiple customers according to the customer attribute information; Calculate the multi-dimensional evaluation parameters of each customer based on the existing product information, transaction information and user product contact behavior information of each customer, wherein the multi-dimensional evaluation parameters include: consumption time interval information, consumption frequency information, consumption amount information, asset information and product contact information; For each group, the customers in each group are sorted according to each dimension of evaluation parameters; based on the customer sorting results of each dimension of evaluation parameters of each group, the scores of the customers in each group in each dimension of evaluation parameters are calculated; The customers in each group are classified according to their scores in each dimension of evaluation parameters.

2. The method according to claim 1, characterized in that Also includes: Process missing values ​​and outliers in the customer data to filter out customer data that does not meet the requirements; Standardize the variables in the processed customer data; The categorized variables in the processed customer data are coded as dummy variables.

3. The method according to claim 2, characterized in that Standardization of variables in processed customer data includes: For each variable, determine the maximum and minimum values ​​of the variable in all customer data after processing; Divide the value of the variable in each customer data after processing by the difference between the maximum and minimum values ​​to obtain the standard value of the variable.

4. The method according to claim 1, characterized in that According to the customer ranking results of each dimension of evaluation parameters in each group, the scores of customers in each group in each dimension of evaluation parameters are calculated, including: Sort the customers in each group according to the order of each dimension evaluation parameter from small to large; According to the customer ranking results of each dimension of the evaluation parameter of each group, calculate the proportion of the ranking position of the customers in each group in the evaluation parameter ranking results to the total number of customers in the group; The proportion of the ranking positions of customers in each group in the evaluation parameter ranking results to the total number of customers in the group is matched with the preset quantile scoring library of the evaluation parameters to determine the scores of customers in each group in each dimension of the evaluation parameters, wherein the preset quantile scoring library of each evaluation parameter records the correlation between the quantile range and the score.

5. The method according to claim 1, characterized in that According to the scores of the customers in each group in each dimension of evaluation parameters, the customers in each group are classified, including: Match the scores of the evaluation parameters of each dimension of the customers in each group with the customer classification library, and use the matched type as the type of the customers in each group; The customer classification library stores the association between customer types and evaluation parameter score ranges.

6. The method according to claim 5, characterized in that Also includes: According to the customer type, the information push strategy is determined by querying the strategy library, wherein the strategy library includes the association between the customer type and the strategy; Push information to customers according to the information push strategy.

7. The method according to claim 5, characterized in that Also includes: Screen out customers of the intended customer type; Input the multi-dimensional evaluation parameters of the screened customers and the product information to be pushed into the pre-established product preference prediction model to obtain the customer's preference for the pushed product information; Push product information to customers based on their preference for the pushed product information; The product preference prediction model is trained based on a training sample set established based on historical response information of multiple customers of a predetermined customer type to push different products, wherein the training sample set includes the customer's multi-dimensional evaluation parameters, product information, and the customer's preference probability for the product; The product preference prediction model includes: a support vector machine module based on a hybrid kernel function and an rank-preserving regression module.

8. A customer classification device, characterized in that: include: an acquisition unit, used to acquire customer data of a plurality of customers within a predetermined time period; A calculation unit is used to calculate the multi-dimensional evaluation parameters of each customer based on the acquired customer data, wherein the customer data includes customer attribute information, asset information, existing product information, transaction information and user product contact behavior information within a predetermined time period, and the multi-dimensional evaluation parameters include: consumption time interval information, consumption frequency information, consumption amount information, asset information and product contact information; A scoring unit, used to calculate the customer's score in each dimension of evaluation parameters according to the customer ranking results according to each dimension of evaluation parameters; The classification unit is used to classify customers according to their scores in each dimension of evaluation parameters.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor of a computer device, the method according to any one of claims 1 to 7 is implemented.

Images