CN115423601A - Method and device for designing online credit product - Google Patents

Abstract

The invention relates to the technical field of financial credit, and particularly provides a method for designing an online credit product, which comprises the following steps: s1, a user credit granting application; s2, wind control and credit granting; s3, rechecking the credit approval system; s4, applying for money; and S5, surface labeling. Compared with the prior art, the method can reduce paper approval data, improve approval timeliness and improve the ability of obtaining and keeping guests through accurate marketing.

Description

Method and device for designing online credit product

Technical Field

The invention relates to the technical field of financial credit, and particularly provides a method and a device for designing an online credit product.

Background

At present, internet finance is rapidly developed and competition in the same industry is continuously intensified, and the situation that on-line channels for customers are deficient, capital operation cost is high and traditional advantages are gradually lost is shown in the banking industry and various financial institutions. In order to provide better financial services for small and medium-sized enterprises in the jurisdiction, widen channels for obtaining customers, accelerate transformation development and provide an online financial solution, online loan products are manufactured by taking enterprise business credit products as an entry point, taking enterprise multi-source multi-dimensional data as a core and taking a product platform as a channel. Meanwhile, on the basis of not changing the product property, the traditional offline real name, offline data collection and offline credit process are online by processing public data, organization data and internal data and combining a novel financial science and technology means, so that the user can complete the process from application to credit without meeting with a customer manager before surface signing. The method has the advantages that user experience is optimized to the maximum extent, the enterprise loan application efficiency is improved, the working cost of a customer manager is reduced, manual intervention in the credit granting process is avoided, and the product coverage rate is improved.

The online transformation of financial products is a channel for simplifying the offline complex auditing process, and the key for determining whether the products are used is how to design a financial product with data support and strong referential significance. The financial wind control credit assessment plays an increasingly obvious role in the field of financial wind control, and how to construct a risk assessment system suitable for small and micro enterprises is particularly important.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides a method for designing an online credit product with strong practicability.

The invention further aims to provide a reasonable-design, safe and applicable online credit product design device.

The technical scheme adopted by the invention for solving the technical problem is as follows:

a design method of on-line credit product comprises the following steps:

s1, a user credit granting application;

s2, wind control and credit granting;

s3, rechecking the credit approval system;

s4, applying for money;

and S5, surface labeling.

Further, in step S1, the method further includes:

s101, real-name authentication:

the real-name authentication comprises three types of personal real-name authentication, enterprise real-name authentication and sponsor real-name and authorization;

s102, information acquisition, image data entry and verification:

firstly, directly using a data index item, seeking data field replacement for data which cannot be acquired in a homogeneous data conversion mode, and finally, collecting the residual data which cannot be acquired and needs to be stored to a user in an image data uploading mode;

s103, information pre-examination:

the data passing the front-end verification is transmitted to an inline credit approval system in the form of a data interface, and a customer manager backs the image data which is not verified in detail and requests the loan application enterprise to upload again.

Further, in step S2, the method further includes:

s201, data access;

s202, processing and developing service indexes;

s203, designing and developing an intelligent wind control module;

and S204, deploying the intelligent wind control module.

Further, in step S201, the data access deploys the intra-system to a data line of the product platform for data transmission, and the wind-controlled data source includes five types, i.e., tax data, personal credit data, enterprise credit data, house property data, and corporate network data.

Further, in step S202, index processing and index development are further included,

the index processing is used for compiling business index derivation and calculation processing logic according to the collected data interface documents and in combination with business backgrounds, in the process of processing the business indexes, the index processing documents are submitted in stages for business experts to review, and the revision of the index processing logic is completed according to review opinions;

the indexes passed by the index development process documents, the programming and development of the service indexes are completed, the index development logic is perfected through the investigation and statistics of the missing values of the abnormal values of the indexes, and the output index values are ensured to meet the quality requirements of the intelligent wind control module.

Further, in step S203, the method further includes:

(1) And combining the data source, and forming a product admission module from the class to which the index belongs, the index definition logic, the admission rule and the data source.

(2) The product anti-fraud module consists of anti-fraud rules, sets rejection conditions according to the service definition of the data indexes, triggers the application of the rejection conditions, and rejects the rejection, wherein the rejection comprises enterprise legal anti-fraud rules and enterprise anti-fraud rules;

(3) Internal enterprise scoring, wherein model construction is carried out by combining an AHP wind control scoring card based on gradient lifting algorithm feature importance ranking with business personnel wind control experience, the inline scoring mainly evaluates the enterprise credit condition from two dimensions of quantitative factors and qualitative factors, the quantitative factors mainly evaluate the financial condition of the enterprise, and the qualitative indexes depend on the subjective evaluation of a customer manager and credit examiners;

(4) And dividing the enterprise into 9 levels according to the enterprise scoring result, wherein the levels are AAA, AA, A, BBB, BB, B, CCC, CC and C, only the BBB level and above clients are allowed to access and examine the pre-authorized credit limit, the authorized credit limit calculation formula of the clients refers to the original internal credit standard, and the tax payment condition of the enterprise in the last two years is considered in the selection of the credit payment base number.

Further, in step S204, after the intelligent wind control module is developed, the model construction result is packaged and deployed in the inline production environment, and after the intelligent wind control module is deployed online, the intelligent wind control module receives the service application from the product platform, matches the wind control data of various data sources, automatically processes the received service application to generate a credit granting result or a credit granting scheme, and feeds the credit granting result or the credit granting scheme back to the bank wind control department.

Further, in step S3, a credit approval scheme including credit approval rating and pricing is generated in the bank, the credit approval system opens a corresponding data interface to realize system data docking, automatically pushes and uploads the required data, and the credit approval system includes the credit approval scheme and the early-stage exhausted data.

Further, in step S4, after the user sees the credit granting plan at the client, the user can click the application for using money to apply for loan; the client prompts paper materials and specification requirements required to be prepared when the user signs on the surface;

in step S5, the face sign is divided into off-line paper data retention, bank account opening and sign-up deposit.

An apparatus for designing an on-line lending product, comprising: at least one memory and at least one processor;

the at least one memory to store a machine readable program;

the at least one processor is configured to invoke the machine readable program to perform a method for designing an online loan product.

Compared with the prior art, the design method and the device for the online credit product have the following outstanding beneficial effects:

(1) The design of the on-line credit product provided by the invention can reduce paper approval data, improve the approval timeliness and improve the customer-obtaining ability through accurate marketing.

(2) The whole process from data collection and input to evaluation result output is completed by a computer algorithm, so that the artificial subjective judgment is reduced, the objectivity of the evaluation result is ensured, and the reduction of the operation risk of the approval process is facilitated.

(3) The method can reduce overfitting and calculation, improves the efficiency of the algorithm to a great extent, further calculates the dimension weight ratio of each index by converting the average score of each index into the initial judgment matrix ratio, and comprehensively calculates to obtain the final scoring model. The influence of human intervention on the model output is reduced, the value attribute of the data is fully activated, and the automation degree of the model operation is further improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

FIG. 1 is a schematic flow chart of a method for designing an online loan product;

FIG. 2 is a schematic diagram of an enterprise credit wind control modeling process in an online credit product design method.

Detailed Description

The present invention will be described in further detail with reference to specific embodiments in order to better understand the technical solutions of the present invention. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

A preferred embodiment is given below:

as shown in fig. 1, the method for designing an online credit product in this embodiment includes the following steps:

s1, a user credit granting application;

the user credit application refers to a process that a user with real name authentication initiates credit application through a client, and after an authorized related unit inquires information of enterprises and legal persons of the user, related credit supplement data and electronic image data are filled in the client and transmitted to an inline service background for storage. When a user initiates a credit application, the system automatically generates a user ID and an order flow, and unique identification of the user and the flow is made in the system.

The method specifically comprises the following steps:

s101, real name authentication

The real-name authentication comprises three types of personal real-name authentication, enterprise real-name authentication and sponsor real-name and authorization.

The personal real-name authentication adopts four-factor authentication of an operator, namely three factors of a name, a mobile phone number and an identity card number are kept consistent. And the human face living body monitoring authentication with higher authentication grade, namely the name, the identity card number and the human face living body detection are consistent.

The enterprise real-name authentication adopts four-factor authentication of enterprise industry and commerce, namely four factors of unified social credit codes, enterprise names, identity card numbers of legal persons and names of legal persons are consistent.

The real-name authorization of the sponsor refers to that the sponsor after real-name authentication sends authorization operation to a legal person in a short message mode, and requests the legal person to authorize the right of borrowing by a legal person represented by other people. The authorization short message is that a short message interface is triggered by a sponsor, a system automatically generates a link containing real-name information of a legal person and sends the link to a mobile phone of the legal person, the legal person needs to be opened in a client side for safety, and the legal person signs an authorization book in an electronic signature mode after passing personal real-name authentication, namely, the authorization of the legal person is completed.

S102, information acquisition, image data entry and verification

Firstly, analyzing and processing internet data, organization data and enterprise data of a loan application enterprise, comparing and analyzing the internet data, the organization data and the enterprise data with a traditional credit extension material, and preferentially searching a data index item which can be directly used; meanwhile, data information which cannot be acquired is analyzed and judged based on data use, and other data field replacement is sought in a homogeneous data conversion mode; and finally, collecting the residual data which cannot be acquired and needs to be stored to the user in a mode of uploading the image data. The collection of image data can be performed with preliminary front-end check, and the data passing through the front-end check is transmitted to the inline system in the form of data interface.

Information acquisition and entry information can be roughly classified into five categories: borrowing willingness documents, basic conditions of enterprises, enterprise operation materials, account opening and bank flow data, other additional data and the like.

S103, information pre-examination

And pre-reviewing information, namely transmitting data which passes through front-end verification to an inline credit approval system in a data interface mode according to the early-stage requirement investigation condition, and backing the image data which is not verified in detail by a customer manager and requiring a credit application enterprise to upload again, so that the condition that the user prepares the data completely and accords with inline compliance audit requirements on the premise of not meeting the customer manager is ensured.

S2, wind control and credit granting;

the intelligent wind control module of the product is packaged and deployed in the in-line wind control system, and comprises four steps of data access, service index processing and development, intelligent wind control module design and development and intelligent wind control module deployment.

Further comprising:

s201, data access;

data access requires deployment of an inline system to a data line of a product development platform for data transmission. The wind control data source comprises five types of tax data, personal credit investigation data, enterprise credit investigation data, real estate data and converged network data. And the access paths of all data sources are completed by the cooperation of a pedestrian party and a product construction platform party, and if necessary, relevant government departments are coordinated, and after compliance desensitization is carried out according to the data compliance requirements, the original path is fed back to obtain a result.

S202, processing and developing service indexes;

1) Index processing: product index processing requires compiling business index derivation and calculation processing logic according to collected data interface documents and in combination with business backgrounds. In the process of processing the service index, submitting the index processing document by stages for the service expert to review, and finishing the revision of the index processing logic according to the review suggestion.

2) Index development: and processing the document according to the index passing the evaluation, and finishing the programming and development of the service index. By checking and counting the loss values of the index abnormal values, the index development logic is perfected, and the output index values are ensured to meet the quality requirements of the intelligent wind control module. The original field used by index development needs to be stored and backed up, so that the tracing of the index processing logic is convenient.

S203, designing and developing an intelligent wind control module;

1) Product access module

The product admission module fully refers to the product demand policy in the row and combines all data sources, and the product admission module is formed from multiple dimensions such as the type of the index, the index definition logic, the admission rule, the data source and the like. As an entry level condition for enterprise loan application, the product access module fully considers the basic information of enterprises and legal persons, the tax condition, the operation behavior and other indexes, such as two years or more of continuous and stable tax payment of the enterprises. Through comparative analysis, indexes in the product access model of the wind control scheme cover all access conditions of loan products of small and micro enterprises in a line, the flow automation of the access auditing of the products in the line is realized, the auditing workload of personnel in a bank is reduced, and the auditing efficiency is improved.

2) Anti-fraud module

The anti-fraud module is the core of the wind control system, and in order to guarantee the effectiveness and the safety of the rule and reduce the risk exposure caused by the error of rule strategy setting and the error of list adding, the wind control system strictly controls the setting and the threshold value of the anti-fraud module. The product anti-fraud module is composed of anti-fraud rules, sets rejection conditions according to the service definition of the data indexes, triggers the application of the rejection conditions, and rejects. Contains enterprise legal anti-fraud rules (individual rejection rules, individual prompt rules) and enterprise anti-fraud rules (enterprise rejection rules, enterprise prompt rules). Calling by an anti-fraud module: and (3) rejecting the client by touching the rule, outputting fraud risk level by touching, carrying out manual approval on the borrowers above a lower risk level by an approval person, and carrying out automatic approval credit granting by the system at low risk and lower risk levels.

3) Internal credit scoring

And (4) grading enterprises in the product, and performing model construction by adopting an improved AHP wind control grading card based on gradient lifting algorithm (XGboost) feature importance ranking in combination with business personnel wind control experience. The inline rating mainly evaluates the credit condition of an enterprise from two dimensions of quantitative factors and qualitative factors, the quantitative factors mainly evaluate the financial condition of the enterprise, and the qualitative indexes depend on the subjective evaluation of a customer manager and credit auditors.

The credit scoring model comprehensively evaluates the credit condition of an enterprise from multiple dimensions, selects an input index according to business experience on the basis of investigating the internal wind control rating, breaks through the traditional mode of determining a judgment matrix by expert experience, and has the advantage that the importance score of each attribute can be relatively directly obtained after a promotion tree is created by adopting an XGboost algorithm. Generally, the importance score measures the value of a feature in the construction of a boosted decision tree in a model. The more an attribute is used to build a decision tree in the model, the higher its importance is relative. The attribute importance is obtained by calculating and sorting each attribute in the data set. The attribute importance is calculated in a single decision tree by the amount that each attribute split point improves the performance metric. The node is responsible for weighting and recording times, namely the larger the improved performance measurement of one attribute on the split point (closer to the root node), the larger the weight value is; the more the promotion tree is selected, the more important the attribute is. Performance metrics Gini purity of the split node is chosen here as a metric. And finally, carrying out weighted summation on results of one attribute in all the lifting trees, then averaging to obtain a feature importance score, and further converting the feature importance average score value into each index to establish an initial judgment matrix. The weight of the index of the layer to the criterion is obtained by calculating the maximum eigenvalue of the judgment matrix and the corresponding orthogonalized eigenvector, and the combined weight of each layer of indexes to the overall target is calculated on the basis, so that the final weight of each dimension index is determined.

4) Quota pricing module

Through the construction of the credit scoring model, the credit scoring of the enterprise is comprehensively evaluated, the enterprise is divided into 9 levels according to the scoring result of the enterprise, wherein the levels are AAA, AA, A, BBB, BB, B, CCC, CC and C, and only BBB level or above customers are allowed to approve the pre-granted credit limit. The credit line calculation formula of the access customer refers to the original credit standard in the line, and the selection of the credit base number considers the tax payment condition of the enterprise in the last two years.

S204, deploying intelligent wind control modules

And after the intelligent wind control module is developed, the model construction result is packaged and deployed in the inline production environment. After the intelligent wind control module is deployed on line, the intelligent wind control module receives service application from a product platform, matches wind control data of various data sources, automatically processes and generates a credit granting result or a credit granting scheme, and feeds back the credit granting result or the credit granting scheme to a bank wind control department.

S3, rechecking the credit approval system;

through a previous intelligent wind control process, a credit approval scheme containing credit approval rating and pricing is generated in a bank, a credit approval system opens a corresponding data interface to realize system data butt joint, required data can be automatically pushed and uploaded, and the credit approval system mainly comprises the credit approval scheme and previous-stage exhausted data (which can be inquired and downloaded from a related data port of a product construction platform and the like) and is input into the credit approval system. The credit approval system can generate a survey analysis report for assisting in completing the credit granting scheme according to the existing mode, wherein the survey analysis report comprises a rating and rating table, an examination and approval opinion table, an interest rate measuring and calculating table and the like, a loan application book and the like. And the wind control manager feeds the credit back to the client through the product platform. After the credit approval system is formally operated and used online, the workload of a customer manager and a wind control manager is greatly reduced.

S4, applying for money;

after seeing the credit granting scheme at the client, the user can click the payment application to apply for loan. The client prompts paper materials and specification requirements (simultaneously, the face-signing notification is carried out on a legal person (or a entrusted manager) and an actual controller in a short message mode) which are required to be prepared when a user faces to sign, the information includes but is not limited to information supplementary recording materials and uploaded image data, the data are guaranteed to be real and complete, the user only needs to face to sign once, and complex offline procedures are simplified.

S5, surface labeling;

the surface label is an offline existing flow and mainly comprises three parts: off-line paper data retention, bank account opening, signing and paying.

1) The face signing is completed by the legal person and the spouse together, and the early authorization and the identity authentication need to be verified;

2) The surface label needs to sign documents such as loan application books, legal guarantee and the like;

3) When the face sign is made, the face sign business personnel can directly search out the borrowing order to accept the face sign by simultaneously marking through the identity card of the legal person or the unified social credit code of the enterprise.

An apparatus for designing an online loan product, comprising: at least one memory and at least one processor;

the at least one memory to store a machine readable program;

the at least one processor is configured to invoke the machine readable program to perform a method for on-line lending product design.

The above embodiments are only specific examples of the present invention, and the scope of the present invention includes but is not limited to the above embodiments, and any suitable changes or substitutions that may be made by one of ordinary skill in the art and in accordance with the method and apparatus for designing an online credit product according to the present invention and the claims thereof should fall within the scope of the present invention.

Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (10)

1. A method for designing an on-line credit product, comprising the steps of:

s1, a user credit granting application;

s2, wind control and credit granting;

s3, rechecking the credit approval system;

s4, applying for money;

and S5, surface labeling.

2. The method for designing an online credit product according to claim 1, further comprising, in step S1:

s101, real-name authentication:

the real-name authentication comprises three types of personal real-name authentication, enterprise real-name authentication and sponsor real-name and authorization;

s102, information acquisition, image data entry and verification:

firstly, directly using a data index item, seeking data field replacement for data which cannot be acquired in a homogeneous data conversion mode, and finally, collecting the residual data which cannot be acquired and needs to be stored to a user in an image data uploading mode;

s103, information pre-examination:

the data passing the front-end verification is transmitted to an inline credit approval system in the form of a data interface, and a customer manager backs the image data which is not verified in detail and requests the loan application enterprise to upload again.

3. The method as claimed in claim 2, further comprising, in step S2:

s201, data access;

s202, processing and developing service indexes;

s203, designing and developing an intelligent wind control module;

and S204, deploying an intelligent wind control module.

4. The method as claimed in claim 3, wherein in step S201, the data access deploys the in-line system to the data line of the product platform for data transmission, and the wind-controlled data source includes five types, i.e. tax data, personal credit data, enterprise credit data, house property data, and corporate network data.

5. The method as claimed in claim 4, wherein the step S202 further comprises index processing and index development,

the index processing is used for compiling business index derivation and calculation processing logic according to the collected data interface documents and in combination with business backgrounds, in the process of processing the business indexes, the index processing documents are submitted in stages for business experts to review, and the revision of the index processing logic is completed according to review opinions;

the indexes passed by the index development process documents, the programming and development of the service indexes are completed, the index development logic is perfected by checking and counting the missing values of the abnormal values of the indexes, and the output index values are ensured to meet the quality requirements of the intelligent wind control module.

6. The method for designing an online credit product according to claim 5, wherein in step S203, the method further comprises:

(1) And combining the data source, and forming a product admission module from the class to which the index belongs, the index definition logic, the admission rule and the data source.

(2) The product anti-fraud module consists of anti-fraud rules, sets rejection conditions according to the service definition of the data index, triggers the application of the rejection conditions, and rejects the rejection conditions, wherein the rejection conditions comprise enterprise legal anti-fraud rules and enterprise anti-fraud rules;

(3) Internal enterprise scoring, wherein model construction is carried out by combining an AHP wind control scoring card based on gradient lifting algorithm feature importance ranking with business personnel wind control experience, the inline scoring mainly evaluates the enterprise credit condition from two dimensions of quantitative factors and qualitative factors, the quantitative factors mainly evaluate the financial condition of the enterprise, and the qualitative indexes depend on the subjective evaluation of a customer manager and credit examiners;

(4) The enterprise is divided into 9 levels according to the enterprise scoring result, wherein the levels are AAA, AA, A, BBB, BB, B, CCC, CC and C, only the BBB level or more clients are allowed to access and approve the pre-credit line, the credit line calculation formula of the allowed clients refers to the internal original credit standard, and the tax payment condition of the enterprise in the last two years is considered in the selection of the credit line number.

7. The method for designing an online credit product according to claim 6, wherein in step S204, after the intelligent wind control module is developed, the model construction result is packaged and deployed in an inline production environment, and after the intelligent wind control module is deployed online, the intelligent wind control module receives a service application from a product platform, matches wind control data of various data sources, automatically processes the service application to generate a credit granting result or a credit granting scheme, and feeds the credit granting result or the credit granting scheme back to the bank wind control department.

8. The method as claimed in claim 7, wherein in step S3, a credit approval system opens corresponding data interface to implement system data docking, automatically pushes and uploads the required data, and the on-line credit approval system includes the credit approval scheme and the previous-stage depletion data input into the credit approval system.

9. The method as claimed in claim 7, wherein in step S4, after the user sees the credit granting plan at the client, the user can click on the application for use to request for loan; the client prompts paper materials and specification requirements required to be prepared when the user signs on the surface;

in step S5, the face sign is divided into off-line paper data retention, bank account opening and sign-up deposit.

10. An apparatus for designing an on-line lending product, comprising: at least one memory and at least one processor;

the at least one memory to store a machine readable program;

the at least one processor configured to invoke the machine readable program to perform the method of any of claims 1 to 9.

Images