CN116127363A

CN116127363A - Customer classification method, apparatus, device, medium, and program product

Info

Publication number: CN116127363A
Application number: CN202310150469.2A
Authority: CN
Inventors: 唐杨; 李雪
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2023-02-13
Filing date: 2023-02-13
Publication date: 2023-05-16

Abstract

The disclosure provides a customer classification method which can be applied to the technical fields of finance, artificial intelligence, cloud platforms and the like. The client classification method comprises the following steps: acquiring an evaluation index set of a customer to be classified, wherein the evaluation index set of the customer to be classified comprises evaluation indexes of different index categories; based on the evaluation indexes of different index categories of the clients to be classified, matching the evaluation indexes with a preset basic index library to obtain index variables of different index categories, wherein the index variables and the evaluation indexes in the preset basic index library are in one-to-many correspondence; calculating an intermediate classification result based on the index variables of the different index categories; and outputting the intermediate classification result as a final classification result when the intermediate classification result meets a preset classification output result. The present disclosure also provides a customer classification apparatus, device, storage medium, and program product.

Description

Customer classification method, apparatus, device, medium, and program product

Technical Field

The present disclosure relates to the technical field of finance, artificial intelligence, cloud platform, and the like, and in particular to a customer classification method, apparatus, device, medium, and program product.

Background

In the financial field, when a client performs business handling, the background classifies the client according to the relevant effective data of the client, and corresponding measures are adopted for different clients corresponding to different index categories under different business scenes so as to strengthen client management.

The effective data related to the clients are numerous and are respectively stored in different systems, and the excessive data volume can cause excessive consumption of system resources and insufficient timeliness of classification if a traditional traversal tracking method is adopted to search the effective data, so that the clients can be identified and identified after the high risk event occurs after the clients are identified.

Disclosure of Invention

In view of the foregoing, the present disclosure provides client classification methods, apparatus, devices, media, and program products that improve client classification efficiency and accuracy.

According to a first aspect of the present disclosure, there is provided a customer classification method comprising: acquiring an evaluation index set of a customer to be classified, wherein the evaluation index set of the customer to be classified comprises evaluation indexes of different index categories; based on the evaluation indexes of different index categories of the clients to be classified, matching the evaluation indexes with a preset basic index library to obtain index variables of different index categories, wherein the index variables and the evaluation indexes in the preset basic index library are in one-to-many correspondence; calculating an intermediate classification result based on the index variables of the different index categories; and outputting the intermediate classification result as a final classification result when the intermediate classification result meets a preset classification output result.

According to an embodiment of the disclosure, the evaluation index includes a first evaluation index, and the matching between the evaluation index based on different index categories of the clients to be classified and a preset basic index library to obtain index variables of different index categories includes: establishing calculation pools corresponding to the index categories of the clients to be classified one by one, wherein the calculation pools comprise index calculation pools and variable pools; pushing the evaluation indexes of the clients to be classified and index variables corresponding to the preset basic index library to the index calculation pool of the corresponding index class, wherein the index variables of the preset basic index library are pushed randomly; calculating the offset between the evaluation index of the customer to be classified and the corresponding evaluation index in the preset basic index library; selecting the evaluation index of the offset preset offset interval as the first evaluation index; and acquiring corresponding index variables in the preset basic index library based on the first evaluation index so as to store the index variables into the variable pool.

According to an embodiment of the disclosure, the index calculation pool includes different priorities, and after the calculation pool corresponding to the index category of the client to be classified is established, the method further includes: calculating a pool matching container based on the variable pool and the index respectively, wherein the container comprises a plurality of containers with performance from high to low, and the matching method of the variable pool comprises the following steps: the variable pool is preferentially matched in the order of performance from high to low based on the container compared with the index calculation pool, and the matching method of the index calculation pool comprises the following steps: and after the variable pool matching is finished, matching the index calculation pool and the container based on the remaining containers, wherein the priority of the index calculation pool is positively correlated with the performance of the container.

According to an embodiment of the present disclosure, wherein different variable pools correspond to different variables Chi Quan, the calculating the intermediate classification result based on the index variables of the different index categories includes: and calculating based on the Chi Quan weight of the variable and the corresponding index variable to obtain an intermediate classification result.

According to an embodiment of the present disclosure, when the intermediate classification result meets a preset classification output result, the outputting the intermediate classification result as a final classification result includes: and under the condition that the evaluation index of the customer to be classified is completely matched with the corresponding index variable in the preset basic index library, taking the final intermediate classification result as the final classification result.

According to an embodiment of the disclosure, when the intermediate classification result meets a preset classification output result, the intermediate classification result is output as a final classification result, and the method further includes: and under the condition that the evaluation index of the customer to be classified is not completely matched with the corresponding index variable in the preset basic index library, when the intermediate classification result is in a preset threshold value, taking the intermediate classification result as the final classification result.

According to an embodiment of the present disclosure, the method for establishing the preset base index library includes: acquiring database building data, wherein the database building data comprises evaluation indexes and index variables of a plurality of index categories, and the evaluation indexes and the index variables are in a many-to-one relationship; and storing the evaluation index and the index variable according to the index category, and establishing a mapping relation between the evaluation index and the index variable.

According to an embodiment of the disclosure, the storing the evaluation index and the index variable according to the index category, and establishing a mapping relationship between the evaluation index and the index variable includes: and converting the evaluation index into a hash value, and establishing a mapping relation between the hash value and the index variable.

According to an embodiment of the present disclosure, the calculating an offset between the evaluation index of the customer to be classified and a corresponding evaluation index in the preset base index library includes: calculating a hash value of the evaluation index of the client to be classified; and calculating the offset based on the hash value of the evaluation index of the client to be classified and the hash value of the evaluation index of the preset basic index library.

According to an embodiment of the present disclosure, the obtaining, based on the first evaluation index, a corresponding index variable in the preset base index library includes: and acquiring a corresponding index variable based on the hash value of the first evaluation index and the mapping relation.

According to an embodiment of the disclosure, the index category includes at least a client information category and a client biometric category, the client biometric category includes at least a facial feature category, the preset basic index library includes at least a client information category library and a client biometric library, and the acquiring the library building data includes: acquiring a facial image for the facial feature class; extracting facial image features based on the facial image; based on the facial image features, outputting facial image variables corresponding to the facial image features through a preset machine learning model, wherein the facial image variables are the index variables under the facial feature categories.

According to an embodiment of the disclosure, the customer biometric library is established based on a cloud storage database.

In a second aspect of the present disclosure, there is provided a customer classification apparatus comprising: the system comprises an index acquisition module, a classification module and a classification module, wherein the index acquisition module is used for acquiring an evaluation index set of a customer to be classified, and the evaluation index set of the customer to be classified comprises evaluation indexes of different index categories; the index matching module is used for matching with a preset basic index library based on the evaluation indexes of different index categories of the clients to be classified to obtain index variables of the different index categories, wherein the index variables and the evaluation indexes in the preset basic index library are in one-to-many correspondence; the intermediate classification result calculation module is used for calculating an intermediate classification result based on the index variables of the different index categories; and the final result output module is used for outputting the intermediate classification result as a final classification result when the intermediate classification result meets a preset classification output result.

According to the embodiment of the disclosure, the evaluation indexes comprise first evaluation indexes, and the index matching module is used for establishing calculation pools corresponding to index categories of the clients to be classified one by one, wherein the calculation pools comprise index calculation pools and variable pools; pushing the evaluation indexes of the clients to be classified and index variables corresponding to the preset basic index library to the index calculation pool of the corresponding index class, wherein the index variables of the preset basic index library are pushed randomly; calculating the offset between the evaluation index of the customer to be classified and the corresponding evaluation index in the preset basic index library; selecting the evaluation index of the offset preset offset interval as the first evaluation index; and acquiring corresponding index variables in the preset basic index library based on the first evaluation index so as to store the index variables into the variable pool.

According to an embodiment of the present disclosure, the index calculation pool includes different priorities, and the index matching module is configured to calculate a pool matching container based on the variable pool and the index, respectively, where the container includes a plurality of containers with performance from high to low, and the matching method of the variable pool includes: the variable pool is preferentially matched in the order of performance from high to low based on the container compared with the index calculation pool, and the matching method of the index calculation pool comprises the following steps: and after the variable pool matching is finished, matching the index calculation pool and the container based on the remaining containers, wherein the priority of the index calculation pool is positively correlated with the performance of the container.

According to the embodiment of the disclosure, different variable pools correspond to different variables Chi Quan, and the intermediate classification result calculation module is configured to calculate based on the variables Chi Quan and the corresponding index variables to obtain an intermediate classification result.

According to an embodiment of the disclosure, the final result output module is configured to, when the evaluation index of the customer to be classified is completely matched with a corresponding index variable in the preset basic index library, take the final intermediate classification result as the final classification result.

According to an embodiment of the disclosure, the final result output module is configured to, when the intermediate classification result is at a preset threshold value and the evaluation index of the customer to be classified is not completely matched with the corresponding index variable in the preset base index library, take the intermediate classification result as the final classification result.

According to an embodiment of the disclosure, the apparatus further includes a database building module configured to obtain database building data, where the database building data includes a plurality of index categories of evaluation indexes and index variables, and the evaluation indexes and the index variables are in a many-to-one relationship; and storing the evaluation index and the index variable according to the index category, and establishing a mapping relation between the evaluation index and the index variable.

According to the embodiment of the disclosure, the database building module is used for converting the evaluation index into a hash value and building a mapping relation between the hash value and the index variable.

According to the embodiment of the disclosure, the index matching module is used for calculating a hash value of an evaluation index of the client to be classified; and calculating the offset based on the hash value of the evaluation index of the client to be classified and the hash value of the evaluation index of the preset basic index library.

According to an embodiment of the disclosure, the index matching module is configured to obtain a corresponding index variable based on the hash value of the first evaluation index and the mapping relationship.

According to an embodiment of the disclosure, the index category includes at least a client information category and a client biometric category, the client biometric category includes at least a facial feature category, and the library building module is configured to obtain a facial image for the facial feature category; extracting facial image features based on the facial image; based on the facial image features, outputting facial image variables corresponding to the facial image features through a preset machine learning model, wherein the facial image variables are the index variables under the facial feature categories.

In a third aspect of the present disclosure, there is provided an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the client classification method described above.

In a fourth aspect of the present disclosure, there is also provided a computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the above-described customer classification method.

In a fifth aspect of the present disclosure, there is also provided a computer program product comprising a computer program which, when executed by a processor, implements the above-described customer classification method.

In the embodiment of the disclosure, the evaluation indexes of the clients to be classified under different categories are obtained and then are matched, so that all data do not need to be traversed, less system processing expenditure is ensured, the classification efficiency of the clients can be improved, and the stability of the system can be ensured particularly in a scene that a large number of clients need to be classified at the same time. And the completion of the completion indexes under each index category is matched, the evaluation indexes under each index category are synthesized, the complete quantity of relevant information of the clients to be classified is not needed, the clients can be classified only by partial relevant data of the clients to be classified, the classification can still be completed under the condition that the data of the clients to be classified are missing, the adaptation scene of the client classification method is enhanced, and the compatibility is improved.

Drawings

The foregoing and other objects, features and advantages of the disclosure will be more apparent from the following description of embodiments of the disclosure with reference to the accompanying drawings, in which:

FIG. 1 schematically illustrates an application scenario diagram of customer classification according to an embodiment of the present disclosure;

FIG. 2 schematically illustrates a flow chart of a customer classification method according to an embodiment of the disclosure;

FIG. 3 schematically illustrates a flow chart of an index variable matching method according to an embodiment of the disclosure;

FIG. 4 schematically illustrates a flow chart of another index variable matching method according to an embodiment of the disclosure;

FIG. 5 schematically illustrates an intermediate classification result calculation method according to an embodiment of the disclosure;

FIG. 6 schematically illustrates a method of establishing a preset base index library according to an embodiment of the present disclosure;

FIG. 7 schematically illustrates a flow chart of an offset calculation method according to an embodiment of the disclosure;

FIG. 8 schematically illustrates a flow chart of an index variable acquisition method according to an embodiment of the present disclosure;

FIG. 9 schematically illustrates a flowchart of another method of establishing a preset base index library in accordance with an embodiment of the present disclosure;

FIG. 10 schematically illustrates a full architectural diagram of a customer sorting method according to an embodiment of the present disclosure;

FIG. 11 schematically illustrates a block diagram of a customer classification device according to an embodiment of the disclosure; and

fig. 12 schematically illustrates a block diagram of an electronic device adapted to implement a customer classification method according to an embodiment of the disclosure.

Detailed Description

Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is only exemplary and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the concepts of the present disclosure.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and/or the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It should be noted that the terms used herein should be construed to have meanings consistent with the context of the present specification and should not be construed in an idealized or overly formal manner.

Where expressions like at least one of "A, B and C, etc. are used, the expressions should generally be interpreted in accordance with the meaning as commonly understood by those skilled in the art (e.g.," a system having at least one of A, B and C "shall include, but not be limited to, a system having a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.).

In the technical scheme of the disclosure, the authorization or consent of the user is obtained before the personal information of the user is obtained or acquired.

Before the present disclosure is disclosed in detail, key technical terms to be used in the present disclosure are described one by one as follows:

a container: a service program.

Hash distribution: hash distribution is a data distribution method based on a hash function.

Parallel processing: in the parallel processing cluster, each node is provided with an independent storage system and an independent memory system, data are divided into the nodes, and each node is mutually connected through a network, cooperatively calculated and used for providing database service as a whole. The MPP has the core advantage of parallel execution of user requests, and converts all IO requests and CPU calculation and the like of a single-node system into multi-node parallel execution.

The effective data related to the clients are numerous and are respectively stored in different systems, and the excessive data volume can cause excessive consumption of system resources and insufficient timeliness of classification if a traditional traversal tracking method is adopted to search the effective data, so that the clients can be identified and identified after the high risk event occurs after the clients are identified. This makes classification of customers a matter of non-timeliness and fullness with active variables (small sample size, inaccurate results).

To solve the technical problems existing in the prior art, an embodiment of the present disclosure provides a client classification method, including: acquiring an evaluation index set of a customer to be classified, wherein the evaluation index set of the customer to be classified comprises evaluation indexes of different index categories; based on the evaluation indexes of different index categories of the clients to be classified, matching the evaluation indexes with a preset basic index library to obtain index variables of different index categories, wherein the index variables and the evaluation indexes in the preset basic index library are in one-to-many correspondence; calculating an intermediate classification result based on the index variables of the different index categories; and outputting the intermediate classification result as a final classification result when the intermediate classification result meets a preset classification output result.

Fig. 1 schematically illustrates an application scenario diagram of customer classification according to an embodiment of the present disclosure.

As shown in fig. 1, an application scenario 100 according to this embodiment may include

terminal devices

101, 102, 103, a network 104, and a server 105. The network 104 is used as a medium to provide communication links between the

terminal devices

101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The user may interact with the server 105 via the network 104 using the

terminal devices

101, 102, 103 to receive or send messages or the like. Various communication client applications, such as shopping class applications, web browser applications, search class applications, instant messaging tools, mailbox clients, social platform software, etc. (by way of example only) may be installed on the

terminal devices

101, 102, 103.

The

terminal devices

101, 102, 103 may be a variety of electronic devices having a display screen and supporting web browsing, including but not limited to smartphones, tablets, laptop and desktop computers, and the like.

The server 105 may be a server providing various services, such as a background management server (by way of example only) providing support for websites browsed by users using the

terminal devices

101, 102, 103. The background management server may analyze and process the received data such as the user request, and feed back the processing result (e.g., the web page, information, or data obtained or generated according to the user request) to the terminal device.

It should be noted that the client classification method provided in the embodiments of the present disclosure may be generally performed by the server 105. Accordingly, the customer classification device provided by the embodiments of the present disclosure may be generally disposed in the server 105. The client classification method provided by the embodiments of the present disclosure may also be performed by a server or a server cluster that is different from the server 105 and is capable of communicating with the

terminal devices

101, 102, 103 and/or the server 105. Accordingly, the client sorting apparatus provided by the embodiments of the present disclosure may also be provided in a server or a server cluster different from the server 105 and capable of communicating with the

terminal devices

101, 102, 103 and/or the server 105.

It should be understood that the number of terminal devices, networks and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

The client classification method of the disclosed embodiment will be described in detail below with reference to fig. 2 to 10 based on the scenario described in fig. 1.

Fig. 2 schematically illustrates a flow chart of a customer classification method according to an embodiment of the disclosure.

As shown in fig. 2, the client classification method of this embodiment includes operations S210 to S240, and the client classification method may be performed by the server 105.

In operation S210, an evaluation index set of the clients to be classified is obtained, the evaluation index set of the clients to be classified including evaluation indexes of different index categories.

In a typical scenario, when a client performs service handling, the background classifies clients needing to handle the service, thereby implementing client management and assisting in service handling. The business transaction scenarios are varied. For example, the business transaction scenario may categorize the customer in an online marketing scenario, thereby assisting in outputting the marketing strategy; for another example, the business transaction scenario may perform risk classification on the customer in the lending scenario, thereby assisting in outputting the wind control policy; also for example, the business transaction scenario may classify the emotion of the customer in a realistic reception scenario, thereby assisting in outputting the reception policy. The business transaction scenario covers aspects of the financial field, and is not described here in detail.

Therefore, relevant data of the customer needs to be collected before, during or after the business transaction of the customer, the data are used for classifying the customer in the background, and the "relevant data" are called an evaluation index set of the customer to be classified. For classification of clients, the class dimensions referred to need to be diversified, so that comprehensive investigation of clients by means of different index classes is required. For the evaluation indexes in the evaluation index set, different evaluation indexes belong to different index categories. For example, when the index category is a financial index category, a plurality of different financial evaluation indexes are included under the financial index category, for example, the financial evaluation indexes may be liability rate, repayment rate, and the like. That is, the acquired clients' evaluation index sets have a plurality of index categories, and each index category includes a plurality of evaluation indexes.

In operation S220, based on the evaluation indexes of the different index categories of the clients to be classified, the evaluation indexes are matched with a preset basic index library, so as to obtain index variables of the different index categories, wherein the index variables and the evaluation indexes in the preset basic index library have a one-to-many correspondence.

Compared with the evaluation index set with only the evaluation indexes of each index class, the preset basic index library stores not only the evaluation indexes of each index class, but also the evaluation variables of different index classes, wherein the index variables are the variables obtained by summarizing a plurality of different evaluation indexes under the index class, and the index variables are preset by research and development personnel based on the rules of the service. Taking the lending scenario as an example, under the financial index category, when the evaluation index liability rate and the repayment rate are respectively 0.3 and 0.2, the index variable of the corresponding financial index category is 0.3 (interval of 0 to 1). Further, when classifying the lending risk category of the customer, it is necessary to integrate different index variables such as the financial index category, the consumption index category, and the background index category. Therefore, in step S220, the index variable applicable to the service scenario in the preset basic index library is matched by the evaluation index of the customer to be classified.

It should be noted that, in the evaluation index set of the clients to be classified obtained in the above operation S210, the index variable is less complete than the index variable in the preset basic index library. Taking a lending scene as an example, under the scene, the classes such as the financial index class, the consumption index class and the background index class of the client need to be matched, so that under the condition that the complete index classes such as the financial index class, the consumption index class and the background index class and the evaluation indexes subordinate to different index classes are stored in a preset basic index library, the client to be classified can only acquire the evaluation indexes under the financial index class and the consumption index class due to data missing. Therefore, when matching is performed, the clients to be classified can only match the corresponding index variable data in the preset basic index library through the financial index category and the consumption index category, so that the index variable corresponding to each index category is obtained.

It should be noted that, the data pre-stored in the preset basic index library does not relate to the customer to be classified. Only the data features of the customer to be classified are matched with the data features of the past people.

In operation S230, an intermediate classification result is calculated based on the index variables of the different index categories.

It will be appreciated that in the background classification engine, rather than calculating only the data of the client a in the same time period, possibly calculating data of tens of thousands of clients simultaneously, further, the obtained index variable does not instantaneously obtain the full amount of index variable, still further, the final classification result is not instantaneously calculated, but the iteration of the intermediate classification result is completed by continuously iterating the intermediate classification result to finally trigger the cut-off condition.

In operation S240, when the intermediate classification result satisfies a preset classification output result, the intermediate classification result is output as a final classification result.

The condition of cut-off is that after all index variables of clients to be classified are matched completely, an intermediate classification result calculated through all index variables is adopted as a final classification result.

The other cut-off condition is that when the index variable of the customer to be classified is still in the matching process, and the intermediate classification result calculated through the acquired index variable is in a preset range (the preset range can be a certain section of a bidirectional closed section or a certain section of a unidirectional open section), the rest matching operation is cut off, and the intermediate classification result in the preset range is taken as a final result. For example, in a lending scenario, where the risk value (i.e., intermediate classification result) exceeds 200, the customer is classified directly as a high risk user without having to continue matching and calculating other index variables.

It is to be understood that, in the above operations S210 to S240, the corresponding customer representation is matched in the preset basic index library by obtaining the representation information of the customer, so as to obtain the label (i.e., classification result) of the representation. It is emphasized that the final classification result of the customer is the need to aggregate different index data. For example, the customer a to be identified is matched through a preset basic index library, so as to obtain index variables of basic data Z (i.e., evaluation indexes and index variables under all index categories of a certain previous customer) in index category 1, index variables of basic data X in index category 2 and index variables of basic data C in index category 3, and then calculate classification results based on these index variables.

In the embodiment of the disclosure, the evaluation indexes of the clients to be classified under different categories are obtained and then are matched, so that all data do not need to be traversed, less system processing expenditure is ensured, the classification efficiency of the clients can be improved, and the stability of the system can be ensured particularly in a scene that a large number of clients need to be classified at the same time. And the completion of the completion indexes under each index category is matched, the evaluation indexes under each index category are synthesized, the complete quantity of relevant information of the clients to be classified is not needed, the clients can be classified only by partial relevant data of the clients to be classified, the classification can still be completed under the condition that the data of the clients to be classified are missing, the adaptation scene of the client classification method is enhanced, and the compatibility is improved. Since the classification of clients needs to ensure high efficiency, at the same time, the background classification device serves in a multi-person scene at the same time. Therefore, the matching of the index variable and the output of the result may be achieved in an efficient interactive parallel computing manner, that is, the above-described operation S220 and the above-described operation S230 may be achieved in an efficient interactive parallel computing manner.

Fig. 3 schematically illustrates a flowchart of an index variable matching method according to an embodiment of the present disclosure.

As shown in fig. 3, the index variable matching method of this embodiment includes operations S310 to S350, and operations S310 to S350 may at least partially perform operation S220 described above.

In operation S310, a calculation pool corresponding to the index category one of the clients to be classified is established, where the calculation pool includes an index calculation pool and a variable pool.

It should be emphasized that in the embodiments of the present disclosure, there are two index categories, one is an index category of the customer to be identified, and the other is an index category of a preset base index library. The established calculation pools are in one-to-one correspondence with the index categories of the clients to be identified, namely, one calculation pool corresponds to one index category, of course, one calculation pool comprises a plurality of sub-pools, specifically, one calculation pool at least comprises an index calculation pool and a variable calculation pool, wherein the number of the corresponding index calculation pool and the variable calculation pool is not limited.

In operation S320, pushing the evaluation index of the customer to be classified and the index variable corresponding to the preset basic index library to the index calculation pool of the corresponding index class, where the index variable of the preset basic index library is pushed randomly.

It should be noted that, in this stage of operation S320, instead of actively acquiring data from the preset base index library, the preset base index library randomly pushes the evaluation index corresponding to the customer to be classified.

In operation S330, an offset between the evaluation index of the customer to be classified and a corresponding evaluation index in the preset base index library is calculated.

According to an embodiment of the disclosure, the evaluation index comprises a first evaluation index.

In operation S340, the evaluation index of the offset preset offset section is selected as the first evaluation index.

The offset may represent similarity between two data, the offset is affected by an algorithm, and the preset offset interval may be that the offset is greater than a certain offset threshold, or that the offset is less than a certain offset threshold

In operation S350, a corresponding index variable in the preset basic index library is obtained based on the first evaluation index, so as to store the index variable in the variable pool.

It is to be understood that, in the above operations S320 to S350, the evaluation indexes of the clients to be classified are continuously matched with a plurality of different evaluation indexes pushed from a preset basic index library in the calculation pool, so as to calculate the offsets of different targets, select the evaluation indexes with smaller offsets and without excessively affecting the final classification result, and find the index variables corresponding to the evaluation indexes after being completely matched with the subordinate evaluation indexes under one evaluation category.

Of course, if the acquired corresponding index variable of a certain customer to be classified reaches a certain threshold value (in this case, all the index variables of the customer to be classified have not been acquired yet). And tilting the resources of the computing pool towards the clients to be classified, preferentially acquiring the residual index variables of the clients to be classified, and further classifying.

Fig. 4 schematically illustrates a flowchart of another index variable matching method according to an embodiment of the present disclosure.

As shown in fig. 4, the index variable matching method of this embodiment processes the above-mentioned operations S310 to S350, and further includes an operation S410, and the operation S410 is performed after the above-mentioned operation S310 and before the above-mentioned operation S320.

According to an embodiment of the disclosure, the metric calculation pool includes different priorities.

In operation S410, a pool matching container is calculated based on the variable pool and the index, respectively, the container including a plurality of containers with performance from high to low, wherein the variable pool matching method includes: the variable pool is preferentially matched in the order of performance from high to low based on the container compared with the index calculation pool, and the matching method of the index calculation pool comprises the following steps: and after the variable pool matching is finished, matching the index calculation pool and the container based on the remaining containers, wherein the priority of the index calculation pool is positively correlated with the performance of the container.

Specifically, in the sub-pools subordinate to the calculation pool, the index calculation pool includes different priorities. The number of priority levels is predefined by the developer, and for example, the priority levels may be set to priority, secondary, and normal, and the same container may be defined as a high-performance container, a medium-performance container, and a low-performance container. Specifically, for the aspect of system resource occupation distribution, each index calculation pool and each variable pool should follow the following matching principle: the variable pool is always matched with a high-performance container, the computing pool is preferentially matched with the high-performance container, the secondary pool and the common pool are used next, the secondary pool can be matched with the high-performance container when the priority pool is idle, and the common pool can be matched with the high-performance container when the priority pool and the secondary pool are idle.

Fig. 5 schematically illustrates an intermediate classification result calculation method according to an embodiment of the present disclosure.

As shown in fig. 5, the intermediate classification result calculation method of this embodiment includes operation S510, and the operation S510 may at least partially perform operation S230 described above.

According to an embodiment of the present disclosure, different variable pools correspond to different variable pool weights.

In operation S510, a calculation is performed based on the variable Chi Quan weight and the corresponding index variable, resulting in an intermediate classification result.

Specifically, the intermediate classification result is equal to the weighted summation of different index variables, and in a specific service scenario, for example, the classification result needs to be weighted summation of the index variable 1, the index variable 2 and the index variable 3, but in the calculation process, the index variable 1 and the index variable 2 are obtained first, but when the index variable 3 is not obtained, the intermediate classification result is calculated by the index variable 1, the index variable 2 and the corresponding weight first, and when the index variable 3 is obtained, the intermediate classification result can be further updated by the index variable 3 and the corresponding weight. It is emphasized that in the embodiments of the present disclosure, by giving weights to the variable pool, the classification method may implement classification under different traffic scenarios by adjusting the weights.

Fig. 6 schematically illustrates a method of establishing a preset base index library according to an embodiment of the present disclosure.

As shown in fig. 6, the method for establishing the preset base index library of this embodiment includes operations S610 to S620.

In operation S610, database construction data including evaluation indexes and index variables of a plurality of index categories is acquired, wherein the evaluation indexes and the index variables are in a many-to-one relationship.

Specifically, the database building data can be established based on the index data of other clients in the past and combined with corresponding index variables, wherein the index variables are preset and can be output by a machine learning model. Taking a lending scene as an example, the database building data at least comprises evaluation indexes and index variables of all index categories of the offending clients, and evaluation indexes and index variables of all index categories of the non-offending clients, wherein a plurality of evaluation indexes of the same index category influence one index variable.

In operation S620, the evaluation index and the index variable are stored according to the index category, and a mapping relationship between the evaluation index and the index variable is established.

Specifically, different libraries can be built according to different index categories, taking a lending scene as an example, a basic risk level library, a basic financial index library, a basic consumption index library and a basic customer background library are built, wherein,

(1) The base risk level library includes customer ID, hash value, behavioral score, and risk level.

(2) The base financial index library includes a customer ID, an asset liability index, and an asset liability index variable.

(3) The base consumption index library includes a customer ID, a credit card consumption index, and a credit card consumption index variable.

(4) The base customer context library includes customer ID, customer age, image layering, occupation, and customer information index variables.

Of course, in terms of data distribution, a hash distribution manner may be adopted, so that one evaluation index corresponds to one hash value, and when calculation is performed in a calculation pool, the hash value is pushed to complete the calculation. For example, the hash distribution of the evaluation index subordinate to the index type 1 is between O-Z, the hash distribution of the evaluation index subordinate to the index type 2 is between H-N, and the hash distribution of the evaluation index subordinate to the index type 3 is between a-G. The index variable values corresponding to the index category 1, the index category 2 and the index category 3 are between 1 and 9. Therefore, the index variable can be found by pushing, calculating and finding the hash value and further finding the mapping relation of the hash distribution.

Fig. 7 schematically illustrates a flowchart of an offset calculation method according to an embodiment of the present disclosure.

As shown in fig. 7, the offset amount calculating method includes operations S710 to S720. The operations S710 to S720 may at least partially execute the operation S330 described above.

In operation S710, a hash value of the evaluation index of the client to be classified is calculated.

In operation S720, the offset is calculated based on the hash value of the evaluation index of the customer to be classified and the hash value of the evaluation index of the preset base index library.

Fig. 8 schematically illustrates a flowchart of an index variable acquisition method according to an embodiment of the present disclosure.

As shown in fig. 8, the index variable obtaining method includes operation S810. This operation S810 may at least partially perform the above-described operation S350.

In operation S810, a corresponding index variable is acquired based on the hash value of the first evaluation index and the mapping relation.

Specifically, the offset may be implemented by calculating a hamming distance between hash values, where the greater the hamming distance is, the smaller the similarity between images is, and therefore, the largest of the hamming distances is taken out of a certain range of offsets, or, in the case where the hamming distance is calculated to be greater than a certain preset threshold in the calculation process, an evaluation index (hash value) corresponding to the hamming distance greater than the preset threshold is taken as the matched evaluation index. And further obtaining corresponding index variables through a plurality of matched evaluation indexes.

Fig. 9 schematically illustrates a flowchart of another method of establishing a preset base index library according to an embodiment of the present disclosure.

As shown in fig. 9, the method for establishing the preset base index library includes operations S910 to S930. The operations S910 to S930 may at least partially perform the operation S610 described above.

According to an embodiment of the disclosure, the index categories include at least a client information category and a client biometric category, the client biometric category includes at least a facial feature category, and the preset base index library includes at least a client information category library and a client biometric library.

It can be understood that the client information category is client information that can be acquired by the background, and based on different service scenarios, the acquired client information is different, and taking a lending scenario as an example, the client information includes financial index information, consumption index information, background information and the like. These data typically take up a small amount of space and are much smaller than the customer biometric categories, just a string of characters or numbers. For biometric features, it generally includes: face image category, fingerprint category, iris category, voiceprint category, etc.

In the process of data preparation in advance, the large amount of evaluation index data of the biological characteristics is considered, and the data is stored in the storage pressure of the cloud alleviation system. Of course, other data with larger data volume can be put into the cloud storage database as well. For data with smaller data volume, the data can be placed into a system database, and the system database is used for storing the data with smaller data volume, for example, the system database can be configured on a centralized storage database.

In operation S910, a face image is acquired for the facial feature class.

In operation S920, a facial image feature is extracted based on the facial image.

In operation S930, based on the facial image features, a facial image variable corresponding to the facial image features is output through a preset machine learning model, wherein the facial image variable is the index variable under the facial feature class.

The preset machine learning model is trained by taking facial image features as model entering data and taking facial image variables as labels.

In a typical scenario, when a customer performs a business transaction (which may be an online marketing scenario, a lending scenario, and a real-world reception scenario (e.g., perceived emotion)), for some reason, fewer customer-related metrics can be obtained. Thus, matching classification can be performed by acquiring a known face image.

It can be understood that, because of matching facial image features, the preset basic index library should be pre-stored with the corresponding image evaluation index. The past data can be processed as follows: data cleaning and layering (1) image feature extraction (2) and establishing a basic index library (3).

For (1) data cleansing and layering, data is layered by time period. For example, the acquired image is subjected to data cleaning and layering according to the time until the image acquisition, the data within 2 years (inclusive) is the strong correlation layer, the data within 2 years (inclusive) to 5 years (inclusive) is the secondary correlation layer, the data within 5 years (inclusive) to 10 years (inclusive) is the weak correlation layer, and the data above 10 years are removed.

And (2) extracting image features, namely extracting the features of the images through a deep convolutional neural network to obtain facial feature evaluation indexes, and associating the facial feature evaluation indexes with preset classifications. Specifically, the categories of the evaluation index include: muscle, texture, color, line, bone, structure, hair, etc., and correspondingly, the corresponding evaluation indexes under each index category are as follows: (1) muscle: thin, thick, soft and hard, and elastic; (2) texture: shallow, messy and somewhat; (3) color: moles, darkness, unification and darkness; (4) line: bending and straightening and proportion; (5) bone: roughness, size, and sharpness; (6) structure: symmetry, shape and ratio; (7) hair: neat, direction, density, thickness, color, gloss, and curl.

And (3) establishing a basic index library, and dividing index categories of muscles, textures, colors, lines, bones, structures, hairs and the like into 7 sub-libraries according to classification grades, wherein the contents in each sub-library are composed of IDs, evaluation indexes and variable values. And performing machine learning on each layer of data in the basic group by adopting a supervised hash algorithm, mapping to obtain hash distribution of all groups of all clients, forming hash values by each hash deficiency, and determining the corresponding relation between classification and the hash values.

The data in the cloud storage can be expressed as follows according to the principle of pool storage: the first layer is a regional layer and corresponds to the first four bits of the client ID; the second layer is a serial number layer, corresponding to the last five bits of the client ID (from 00001 to 99999); the bottom layer is a fragment layer and corresponds to the extraction layer (index data); and finally, packaging and storing all azimuth fragments of the same client ID+ extraction layer.

Fig. 10 schematically illustrates a full architecture diagram of a customer sorting method according to an embodiment of the present disclosure.

As shown in fig. 10, the system database includes a sub-library 1, a sub-library 2, a sub-library 3, a sub-library 4 and a sub-library 5 (i.e. a basic medium library, a liability library, a customer information library, a consumption library and an asset library in fig. 10), and is a non-cloud database, and the corresponding cloud database is stored by a hierarchical index mode and includes different fragment layers, sequence layers and regional layers.

In the matching calculation process of the evaluation instruction, corresponding evaluation indexes are pushed by a system database and a cloud storage database, the deviation value between two data is calculated by the calculation pools, and index variables in the system database and the cloud storage database are actively acquired and stored in the variable pools under the condition that the calculated deviation value meets the condition, so that a result is output. When the calculated result in the variable pool accords with a preset interval, the variable pool initiative request index calculation pool acquires other evaluation indexes under the ID of the client to be classified.

Based on the client classification method, the disclosure also provides a client classification device. The device will be described in detail below with reference to fig. 11.

Fig. 11 schematically illustrates a block diagram of a client sorting apparatus according to an embodiment of the present disclosure.

As shown in fig. 11, the client sorting apparatus 1100 of this embodiment includes an index acquisition module 1110, an index matching module 1120, an intermediate sorting result calculation module 1130, and a final result output module 1140.

The index obtaining module 1110 is configured to obtain an evaluation index set of a customer to be classified, where the evaluation index set of the customer to be classified includes evaluation indexes of different index categories. In an embodiment, the index obtaining module 1110 may be used to perform the operation S210 described above, which is not described herein.

The index matching module 1120 is configured to match with a preset basic index library based on the evaluation indexes of different index categories of the clients to be classified, so as to obtain index variables of different index categories, where the index variables and the evaluation indexes in the preset basic index library have a one-to-many correspondence. In an embodiment, the index matching module 1120 may be used to perform the operation S220 described above, which is not described herein.

The intermediate classification result calculation module 1130 is configured to calculate an intermediate classification result based on the index variables of the different index categories. In an embodiment, the intermediate classification result calculation module 1130 may be configured to perform the operation S230 described above, which is not described herein.

The final result output module 1140 is configured to output the intermediate classification result as a final classification result when the intermediate classification result meets a preset classification output result. In an embodiment, the final result output module 1140 may be used to perform the operation S240 described above, which is not described herein.

In the embodiment of the disclosure, the evaluation indexes of the clients to be classified under different categories are obtained and then are matched, so that all data do not need to be traversed, less system processing expenditure is ensured, the classification efficiency of the clients can be improved, and the stability of the system can be ensured particularly in a scene that a large number of clients need to be classified at the same time. And the completion of the completion indexes under each index category is matched, the evaluation indexes under each index category are synthesized, the complete quantity of relevant information of the clients to be classified is not needed, the clients can be classified only by partial relevant data of the clients to be classified, the classification can still be completed under the condition that the data of the clients to be classified are missing, the adaptation scene of the client classification method is enhanced, and the compatibility is improved. According to the embodiment of the disclosure, the evaluation indexes comprise first evaluation indexes, and the index matching module is used for establishing calculation pools corresponding to index categories of the clients to be classified one by one, wherein the calculation pools comprise index calculation pools and variable pools; pushing the evaluation indexes of the clients to be classified and index variables corresponding to the preset basic index library to the index calculation pool of the corresponding index class, wherein the index variables of the preset basic index library are pushed randomly; calculating the offset between the evaluation index of the customer to be classified and the corresponding evaluation index in the preset basic index library; selecting the evaluation index of the offset preset offset interval as the first evaluation index; and acquiring corresponding index variables in the preset basic index library based on the first evaluation index so as to store the index variables into the variable pool.

Any of the plurality of the target acquisition module 1110, the index matching module 1120, the intermediate classification result calculation module 1130, and the final result output module 1140 may be combined in one module to be implemented, or any of the plurality of modules may be split into a plurality of modules, according to an embodiment of the present disclosure. Alternatively, at least some of the functionality of one or more of the modules may be combined with at least some of the functionality of other modules and implemented in one module. According to embodiments of the present disclosure, at least one of the target acquisition module 1110, the index matching module 1120, the intermediate classification result calculation module 1130, and the final result output module 1140 may be implemented at least in part as hardware circuitry, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented in hardware or firmware in any other reasonable manner of integrating or packaging the circuitry, or in any one of or a suitable combination of any of the three implementations of software, hardware, and firmware. Alternatively, at least one of the label acquisition module 1110, the index matching module 1120, the intermediate classification result calculation module 1130, and the final result output module 1140 may be at least partially implemented as a computer program module, which when executed, may perform the corresponding functions.

As shown in fig. 12, an electronic device 1200 according to an embodiment of the present disclosure includes a processor 1201, which can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 1202 or a program loaded from a storage section 1208 into a Random Access Memory (RAM) 1203. The processor 1201 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or an associated chipset and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), or the like. Processor 1201 may also include on-board memory for caching purposes. The processor 1201 may include a single processing unit or multiple processing units for performing the different actions of the method flows according to embodiments of the disclosure.

In the RAM1203, various programs and data required for the operation of the electronic apparatus 1200 are stored. The processor 1201, the ROM 1202, and the RAM1203 are connected to each other through a bus 1204. The processor 1201 performs various operations of the method flow according to the embodiments of the present disclosure by executing programs in the ROM 1202 and/or RAM 1203. Note that the program may be stored in one or more memories other than the ROM 1202 and the RAM 1203. The processor 1201 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in the one or more memories.

According to an embodiment of the disclosure, the electronic device 1200 may also include an input/output (I/O) interface 1205, the input/output (I/O) interface 1205 also being connected to the bus 1204. The electronic device 1200 may also include one or more of the following components connected to the I/O interface 1205: an input section 1206 including a keyboard, a mouse, and the like; an output portion 1207 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, a speaker, and the like; a storage section 1208 including a hard disk or the like; and a communication section 1209 including a network interface card such as a LAN card, a modem, or the like. The communication section 1209 performs communication processing via a network such as the internet. The drive 1210 is also connected to the I/O interface 1205 as needed. A removable medium 1211 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is installed as needed on the drive 1210 so that a computer program read out therefrom is installed into the storage section 1208 as needed.

The present disclosure also provides a computer-readable storage medium that may be embodied in the apparatus/device/system described in the above embodiments; or may exist alone without being assembled into the apparatus/device/system. The computer-readable storage medium carries one or more programs which, when executed, implement methods in accordance with embodiments of the present disclosure.

According to embodiments of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, which may include, for example, but is not limited to: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. For example, according to embodiments of the present disclosure, the computer-readable storage medium may include the ROM 1202 and/or the RAM 1203 and/or one or more memories other than the ROM 1202 and the RAM 1203 described above.

Embodiments of the present disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowcharts. The program code, when executed in a computer system, causes the computer system to perform the methods provided by embodiments of the present disclosure.

The above-described functions defined in the system/apparatus of the embodiments of the present disclosure are performed when the computer program is executed by the processor 1201. The systems, apparatus, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the disclosure.

In one embodiment, the computer program may be based on a tangible storage medium such as an optical storage device, a magnetic storage device, or the like. In another embodiment, the computer program can also be transmitted, distributed over a network medium in the form of signals, and downloaded and installed via a communication portion 1209, and/or from a removable medium 1211. The computer program may include program code that may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.

In such an embodiment, the computer program can be downloaded and installed from a network via the communication portion 1209, and/or installed from the removable media 1211. The above-described functions defined in the system of the embodiments of the present disclosure are performed when the computer program is executed by the processor 1201. The systems, devices, apparatus, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the disclosure.

According to embodiments of the present disclosure, program code for performing computer programs provided by embodiments of the present disclosure may be written in any combination of one or more programming languages, and in particular, such computer programs may be implemented in high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. Programming languages include, but are not limited to, such as Java, c++, python, "C" or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

Those skilled in the art will appreciate that the features recited in the various embodiments of the disclosure and/or in the claims may be provided in a variety of combinations and/or combinations, even if such combinations or combinations are not explicitly recited in the disclosure. In particular, the features recited in the various embodiments of the present disclosure and/or the claims may be variously combined and/or combined without departing from the spirit and teachings of the present disclosure. All such combinations and/or combinations fall within the scope of the present disclosure.

The embodiments of the present disclosure are described above. However, these examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the embodiments cannot be used advantageously in combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be made by those skilled in the art without departing from the scope of the disclosure, and such alternatives and modifications are intended to fall within the scope of the disclosure.

Claims

1. A method of customer classification, comprising:

acquiring an evaluation index set of a customer to be classified, wherein the evaluation index set of the customer to be classified comprises evaluation indexes of different index categories;

Based on the evaluation indexes of different index categories of the clients to be classified, matching the evaluation indexes with a preset basic index library to obtain index variables of different index categories, wherein the index variables and the evaluation indexes in the preset basic index library are in one-to-many correspondence;

calculating an intermediate classification result based on the index variables of the different index categories; and

and when the intermediate classification result meets a preset classification output result, outputting the intermediate classification result as a final classification result.

2. The method of claim 1, wherein the evaluation index comprises a first evaluation index,

the method for obtaining the index variable of the different index categories based on the evaluation index of the different index categories of the clients to be classified is matched with a preset basic index library and comprises the following steps:

establishing calculation pools corresponding to the index categories of the clients to be classified one by one, wherein the calculation pools comprise index calculation pools and variable pools;

pushing the evaluation indexes of the clients to be classified and index variables corresponding to the preset basic index library to the index calculation pool of the corresponding index class, wherein the index variables of the preset basic index library are pushed randomly;

Calculating the offset between the evaluation index of the customer to be classified and the corresponding evaluation index in the preset basic index library;

selecting the evaluation index of the offset preset offset interval as the first evaluation index; and

and acquiring corresponding index variables in the preset basic index library based on the first evaluation index so as to store the index variables into the variable pool.

3. The method of claim 2, wherein the metric calculation pool includes different priorities,

after the calculation pools corresponding to the index categories of the clients to be classified one by one are established, the method further comprises the following steps:

calculating pool matching containers based on the variable pool and the index, respectively, the containers comprising a plurality of containers with performance from high to low,

the variable pool matching method comprises the following steps: preferentially matching the variable pool in order of performance from high to low based on the container compared to the index calculation pool,

the matching method of the index calculation pool comprises the following steps: and after the variable pool matching is finished, matching the index calculation pool and the container based on the remaining containers, wherein the priority of the index calculation pool is positively correlated with the performance of the container.

4. The method of claim 3, wherein different variable pools correspond to different variables Chi Quan,

the calculating the intermediate classification result based on the index variables of the different index categories comprises the following steps:

and calculating based on the Chi Quan weight of the variable and the corresponding index variable to obtain an intermediate classification result.

5. The method according to claim 2, wherein outputting the intermediate classification result as a final classification result when the intermediate classification result satisfies a preset classification output result, comprises:

and under the condition that the evaluation index of the customer to be classified is completely matched with the corresponding index variable in the preset basic index library, taking the final intermediate classification result as the final classification result.

6. The method of claim 5, wherein outputting the intermediate classification result as a final classification result when the intermediate classification result satisfies a preset classification output result, further comprises:

and under the condition that the evaluation index of the customer to be classified is not completely matched with the corresponding index variable in the preset basic index library, when the intermediate classification result is in a preset threshold value, taking the intermediate classification result as the final classification result.

7. The method according to any one of claims 2 to 6, wherein the method for establishing the preset basic index library comprises the following steps:

acquiring database building data, wherein the database building data comprises evaluation indexes and index variables of a plurality of index categories, and the evaluation indexes and the index variables are in a many-to-one relationship; and

and storing the evaluation index and the index variable according to the index category, and establishing a mapping relation between the evaluation index and the index variable.

8. The method of claim 7, wherein the storing the evaluation index and the index variable according to the index category and establishing the mapping relationship between the evaluation index and the index variable comprises:

and converting the evaluation index into a hash value, and establishing a mapping relation between the hash value and the index variable.

9. The method of claim 8, wherein the calculating an offset between the evaluation index of the customer to be classified and a corresponding evaluation index in the preset base index library comprises:

calculating a hash value of the evaluation index of the client to be classified; and

and calculating the offset based on the hash value of the evaluation index of the client to be classified and the hash value of the evaluation index of the preset basic index library.

10. The method of claim 9, wherein the obtaining, based on the first evaluation index, a corresponding index variable in the preset base index library, comprises:

and acquiring a corresponding index variable based on the hash value of the first evaluation index and the mapping relation.

11. The method of claim 7, wherein the index categories include at least a customer information category and a customer biometric category, the customer biometric category includes at least a facial feature category, the pre-set base index library includes at least a customer information category library and a customer biometric library,

the obtaining database building data comprises the following steps:

acquiring a facial image for the facial feature class;

extracting facial image features based on the facial image;

based on the facial image features, outputting facial image variables corresponding to the facial image features through a preset machine learning model, wherein the facial image variables are the index variables under the facial feature categories.

12. The method of claim 11, wherein the customer biometric library is established based on a cloud storage database.

13. A customer classification device comprising:

The system comprises an index acquisition module, a classification module and a classification module, wherein the index acquisition module is used for acquiring an evaluation index set of a customer to be classified, and the evaluation index set of the customer to be classified comprises evaluation indexes of different index categories;

the index matching module is used for matching with a preset basic index library based on the evaluation indexes of different index categories of the clients to be classified to obtain index variables of the different index categories, wherein the index variables and the evaluation indexes in the preset basic index library are in one-to-many correspondence;

the intermediate classification result calculation module is used for calculating an intermediate classification result based on the index variables of the different index categories; and

and the final result output module is used for outputting the intermediate classification result as a final classification result when the intermediate classification result meets a preset classification output result.

14. An electronic device, comprising:

one or more processors;

storage means for storing one or more programs,

wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of any of claims 1-12.

15. A computer readable storage medium having stored thereon executable instructions which, when executed by a processor, cause the processor to perform the method according to any of claims 1 to 12.

16. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 12.