TWM599962U - System for intelligently processing loan application - Google Patents

Abstract

A system for intelligently processing loan application is provided. is the system includes a main server that operates an intelligent service platform. The intelligent service platform employs a machine-learning method to learn and train several intelligent models according to external or internal data of a financial institution. The intelligent models implement an unmanned service for fast loan review and appropriation that can efficiently control risks of money laundering and credit of customers in order to provide an appropriate loan pricing and also ensure bank profit. When a loan applicant applies a loan application to a financial institution, a loan program can be approved by using a blacklist screening model to determine if the applicant is in a blacklist, a credit scoring model to obtain a credit inquiry result, and a value-undertaking model to determine an undertaking risk. After that, the system also assists the customer to process identity verification and appropriation after confirming the loan program.

Description

Smart loan review system

This creation relates to a technology of financial loan processing, especially an intelligent loan review system that checks loan applications by establishing an intelligent model.

In the current technology, when someone applies for a loan in a bank or online, the bank reviewer needs to obtain various information about the lender, including basic information, historical transaction information, credit information, including financial and salary certificates, and even It is also necessary to confirm the non-blacklist through the Money Laundering Prevention and Control Center, and to pass a certain amount of loan request after manual review and judgment.

The general loan process requires filling out a complicated loan application form. After completing the loan verification, the loan applicant’s loan limit, interest rate, repayment time and repayment plan are determined, and then the guarantee is carried out. The bank confirms that the lender or guarantor himself signs The procedure of borrowing money or guarantee contract, this part is also a procedure that takes time. After guaranteeing, the bank will allocate funds after checking various documents.

The current loan verification method only uses general-purpose computers to query the lender’s information and past records. The loan verification process is mostly manual, which is not easy to achieve the purpose of rapid loan verification, or it needs to be quickly approved under strict conditions. , Otherwise it is easy to make mistakes, it is possible to expand the risk of nuclear lenders (such as banks).

The specification discloses a smart loan review system that provides users with a fast and efficient loan application process through a smart service platform running in a main server.

When a loan applicant submits a loan application request to a financial institution, in the financial institution, after identification, a blacklist screening model can be used to determine whether the loan applicant belongs to the money laundering blacklist. After that, the applicant can obtain the loan application based on the identity of the loan applicant. The credit information of the lender in the financial institution, joint information, basic information of the applicant, transaction information, and digital browsing information, etc., are entered into a credit scoring model to obtain a credit investigation result. Then, enter the loan application data of the lender, the cost parameters of the financial institution, basic information, and the transaction data of various financial products into an undertaking value model, which can execute the undertaking value judgment and set a pricing table for the loan applicant to determine a The loan plan can be funded after online guarantee.

Further, the blacklist screening model is an intelligent model established by machine learning based on internal data of financial institutions and a blacklist of a money laundering prevention center. The credit scoring model is based on the credit information of financial institutions by machine learning. , Transaction data, personal historical data, and the intelligent model established by the joint solicitation data obtained by a joint credit investigation center, and the undertaking value model is based on the data provided by the financial institution to assess the value of undertaking and the loan history using machine learning methods Data and results, as well as intelligent models established by financial institutions’ cost and risk information.

Preferably, when the loan applicant performs the guarantee, the financial institution can confirm the loan details and conditions with the loan applicant through a communication method, that is, complete the guarantee, and then determine the loan plan including the loan amount, interest rate, time, and repayment plan And the terms of breach of contract.

According to the embodiment of the smart loan review system, the system includes a main server, which runs a smart service platform, is equipped with smart check modules, credit scoring modules, and value evaluation modules, which are learned and trained by machine learning methods. The blacklist screening model, credit scoring model, and commitment value model are developed.

The smart loan review system also includes a risk database. When the verification module establishes a blacklist screening model, it can connect to this risk database and an external money laundering prevention center to obtain the blacklist inside and outside the financial institution. Including a collection review system, which records the loan applicant's joint collection information, basic information of the loan applicant, historical transaction data and digital browsing data. Including a one-to-one guarantee system, which is used to implement a guarantee against a loan applicant and a one-account system to implement appropriations.

When a user applies for a loan, the main server provides a user interface through the Internet for the user to apply for a loan. As described in the embodiment, in a financial institution, after identification, the blacklist screening model can determine the loan application Whether a person belongs to the money laundering blacklist, the credit scoring model can be used to obtain the credit investigation result, and then the loan plan can be determined by the value judgment.

In order to further understand the features and technical content of the present invention, please refer to the following detailed descriptions and drawings about the present invention. However, the drawings provided are only for reference and explanation and are not used to limit the present invention.

10: Internet

100: main server

105: Intelligent Service Platform

151: Smart Check Module

152: Credit Scoring Module

153: Valuation Module

101: Web server

103: Database

107: Examination system

108: Protection System

109: Accounting System

110: Risk Database

14: Web interface

15: User

16: Money Laundering Prevention Center

18: Joint Credit Information Center

201: Personal Credit Application Page

203: Risk Check

205: Issue Check Joint Sign

207: Model Quotation

209: choose plan

211: Send a newsletter

213: online guarantee

215: Appropriation

301: Cardholder Credit Application Page

303: choose plan

305: Announcement

307: Risk Check

309: Model Quotation

311: Send a newsletter

313: online guarantee

315: Appropriation

41: Database

42: Blacklist screening model

43: Credit Scoring Model

44: Undertaking Value Model

Steps S401~S411: Smart loan review process

Steps S501~S513: Smart loan review process

Steps S601~S611: Smart loan review process

Figure 1 shows an embodiment diagram of the architecture of the smart loan review system; Figure 2 shows an embodiment diagram of the process of realizing personal credit with the smart loan review system; Figure 3 shows an embodiment diagram of the process of realizing card credit by the smart loan review system Figure 4 shows one of the flowcharts of the embodiment of the smart loan review method; Figure 5 shows the second flowchart of the embodiment of the smart loan review method; and Figure 6 shows the smart quotation in the method executed by the smart loan review system Example flowchart.

The following are specific specific examples to illustrate the implementation of this creation, and those skilled in the art can understand the advantages and effects of this creation from the content disclosed in this specification. This creation can be Implemented or applied through other different specific embodiments, various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the concept of this creation. In addition, the drawings in this creation are merely schematic illustrations, and are not depicted in actual size, and are stated in advance. The following embodiments will further describe the related technical content of this creation in detail, but the disclosed content is not intended to limit the protection scope of this creation.

It should be understood that although terms such as “first”, “second”, and “third” may be used herein to describe various elements or signals, these elements or signals should not be limited by these terms. These terms are mainly used to distinguish one element from another, or one signal from another signal. In addition, the term "or" used in this document may include any one or a combination of more of the associated listed items depending on the actual situation.

The manual discloses an intelligent loan review system that provides quick loan application and loan issuance solutions. According to loan requests, in addition to the traditional time-consuming and labor-intensive procedures for filling out loan applications, guarantees and appropriations, in addition to using internal financial institutions In addition to data, artificial intelligence technology is introduced, and a model for reviewing user credit loans is established through data analysis in the way of machine learning to automate and accelerate the process of verification, credit scoring and value evaluation in order to control customers’ money laundering risk and credit Risks, and provide appropriate loan pricing to ensure that the bank is profitable, so that it can effectively save manpower and time, and propose a quick loan approval plan.

Referring to the structural embodiment diagram of the smart loan review system shown in Figure 1, the smart model used in the system can be used in the anti-money laundering solution (Anti-Money Laundering, AML), which is generally required by domestic and foreign money laundering prevention centers Use analysis tools to continuously and continuously update and expand its monitoring list. When evaluating user loan applications, obtain domestic and foreign money laundering lists (blacklists) through the Money Laundering Prevention and Control Center, and check batches every day.

According to the proposed embodiment of the smart loan review system, as shown in Figure 1, several smart modules are implemented by software in the main server 100 to establish various evaluation modules required by the system. Among them, an intelligent service platform 105 is proposed, which is equipped with an intelligent checking module 151 realized by a software program, so that a blacklist screening model can be established, which is different from the conventional technology that only performs blacklist comparison to screen users. Instead, it uses machine learning methods to establish a screening list model. At the beginning, you can use the money laundering prevention center's list and screening methods, and then use machine learning methods to learn how to filter out hidden blacklists under various conditions. After establishing the screening model, Can continue to learn the effective loan and repayment records in specific institutions to continuously update the blacklist screening model. For example, when the applicant is not a blacklisted person, it will still be judged as a hidden blacklist based on his transaction records and behavior data; the blacklist screening model can also avoid the problem of incorrect judgment caused by a strict screening mechanism For example, if there are people with the same name and surname, but they are not on the blacklist, the related intelligent technology can eliminate errors or excessive screening errors based on their transaction records and behavior data.

According to the displayed smart loan review system, a credit scoring model is established in the smart service platform 105 using the credit scoring module 152 to review the qualifications of users (applicants). The general review is mainly based on the applicant’s financial resources, Credit status is reviewed, while the intelligent loan review system uses machine learning methods based on external data (such as the joint credit information provided by the joint credit center), individual joint scoring, internal (a bank credit information) transaction data, and personal history The data and various surrounding data are trained by machine learning methods to form a credit scoring model, which becomes a model that can review the qualifications of loan applicants, can effectively evaluate the loan risk of the nuclear lender (such as banks), and can quickly collect data based on various data of loan applicants. Trial.

The intelligent loan review system establishes an undertaking value model through the value evaluation module 153 in the intelligent service platform 105. During the loan review process, the undertaking value is used to confirm the cost (personnel cost, capital) of the lender (such as a bank) Cost) and interest rate to determine the loan amount, interest rate, repayment time and repayment plan of the lender (loan applicant), plus default clauses. Correspondingly, the intelligent loan review system uses machine learning algorithms (such as logistic regression, extreme gradient boosting, deep learning, etc.), various data to assess the value of commitments based on the knowledge provided by financial institutions, plus various The historical data and results (repayment status) of the nuclear loan, plus the cost and risk information of the nuclear loan (such as a specific financial institution), learn and train the undertaking value model. The system must continue to update the undertaking value model based on the new data (cost data and interest rate of borrowing). According to the data of the loan applicant, it will also evaluate the cost and interest rate of the lender and determine the loan applicant’s loan. Amount, interest rate, repayment time and repayment plan, as well as default clauses.

According to the system architecture embodiment diagram shown in Figure 1, the smart loan review system proposes a main server 100, which can be set up at the creditor, such as banks and financial institutions, in which the smart service platform 105 runs to realize the establishment of verification and credit scoring. The smart module of the value evaluation model, as shown in the figure, the smart service platform 105, which implements several smart check modules 151, credit scoring module 152, and value evaluation module 153. Each smart module is a machine The learning method learns and trains multiple intelligent models for loan review based on various data, such as the above-mentioned blacklist screening model, credit scoring model, and commitment value model.

According to an embodiment, the main server 10 provides terminal users 15 through a network 10 to apply for a loan through a user interface (such as a web interface 14). In this example, the web interface 14 can cover various websites that can interface with the main server 10. The software interface of the service provided by the server 101, such as the user interface initiated by a specific software program, allows the user 15 to fill in the information required for the loan application through the fields, including personal information, loan application requirements, and financial certification Etc., so that the main server 10 of the system can establish a database 103 of relevant loan applicants.

The intelligent service platform 105 runs in the main server 100. Once the big data analysis and intelligent calculation according to a specific financial institution are performed through the machine learning algorithm, as described in the above embodiment, a blacklist screening model is established through the intelligent check module 151, This is a model for performing blacklist screening. It is different from the traditional text comparison method. In addition to connecting to the risk database 110 internally, it can also be connected to the money laundering prevention center 16 to directly obtain blacklists from inside and outside financial institutions. The internal blacklist includes the recorded network (based on IP address) blacklist, bank customer blacklist, etc., through this blacklist screening model Type, you can compare the Chinese and English according to the name of the loan applicant, and you can also blacklist screening because of learning various name expressions, combination change results, and various personal information and historical data of the loan applicant, which can avoid excessive screening ( Same as the first name and last name), and improve the accuracy (prevent name change).

The intelligent service platform 105 establishes a credit scoring model through the credit scoring module 152. Through the credit scoring model, the external connection joint credit center 18 obtains real-time joint collection data, and the internal connection collection examination system 107 records the joint collection of the applicant Data, basic data of applicants, historical transaction data, and digital browsing data, etc., are used to input the credit scoring model to obtain credit investigation results.

After the blacklist screening model and credit scoring model proposed by the intelligent loan review system are used to investigate the credit of the applicant, the system then integrates customer information to comprehensively evaluate the risk, which is based on the value evaluation module 15 of the intelligent service platform 105 The undertaking value model established by data analysis and training on the cost parameters proposed by financial institutions, loan applicant information, and various loan plans, through the undertaking value model can more comprehensively measure customer risk and value, and find the point of profit and loss. In order to be able to accurately price based on the expected benefits of financial institutions, including the loan limit and repayment interest rate that can be provided to the applicant, a pricing table can be set based on the comprehensive consideration of customer risk and value. The undertaking value model replaces the traditional manual method of verifying loans. Quotations for quota, interest rate and period can be given according to the individual conditions of each applicant. Once the applicant confirms that the loan conditions are correct, the loan will be allocated in a timely manner so that customers can quickly obtain funds.

Once several intelligent models established by the intelligent service platform 105 generate personalized loan conditions for the loan applicant, the subsequent procedures include confirming with the loan applicant whether they agree to the current loan conditions through the internal guarantee system 108 of the financial institution, and related The information of is synchronously transmitted to the collection review system 107 and the guarantee system 108. After the guarantee is completed, the accounting system 109 will be linked to execute the appropriation. Online guarantee can be carried out. The main server 100 interacts with the user 15 through the web interface 14. The loan application process can be presented to the user 15 through the user interface. The information on whether the loan application is successful or not is also through the web interface. 14 is delivered to user 15.

Through the disclosed intelligent loan review system, personal credit processing can be realized, and the corresponding An example diagram of the related process is shown in Figure 2.

According to the personal credit processing procedure, the user enters the loan application processing procedure provided by the lender (such as a bank and other financial institutions) through the one-person credit application page 201. According to the steps, you can start with the user’s identity authentication. After logging in the system, apply The lender needs to provide the loan application documents, and the lender must obtain the basic information, transaction data, joint solicitation information and digital browsing behavior of the lender.

First of all, if the loan applicant is an existing customer of the financial institution itself, in the step of risk check 203, it can be from inside the bank (as shown in Figure 1 as shown in the collection and review system 107) or from outside (as shown in Figure 1 for money laundering prevention and control). The blacklist proposed by Center 16) is used for screening, and a blacklist screening model is established through a machine learning algorithm for blacklist screening, which can effectively avoid manual screening errors and improve efficiency.

Then proceed to issue check joint 205. This part uses the credit scoring model proposed by the intelligent service platform to perform credit scoring. The input data includes basic information such as the age, gender and financial proof of the applicant, credit card consumption, foreign exchange deposit transactions at the counter, Transaction data such as various financial product transactions, loan balances, recent loan application records, credit card credit behavior and other joint collection data, as well as digital browsing behaviors generated in various forums, social networking sites, etc., these data are obtained through credit scoring models for review result.

Then in the model quotation 207 process, the undertaking value model in the intelligent service platform is used. The input data includes the cost parameters of the financial institution that provides the loan, including hardware costs, labor costs, operating costs, etc., and also includes information about the applicant The basic data, loan application data, and various records of transactions with financial institutions, such as loan records, credit card and foreign exchange records, etc., through the undertaking value model to generate a pricing table for this loan applicant, provide option 209, and pass Send the warranty newsletter 211 for online warranty 213. Once the warranty is completed, you can allocate 215. It is worth mentioning that in this personal credit processing process, if any of the checkpoints is not passed, it means that it has not met the requirements for rapid review, and it can be converted to a manual review process.

Fig. 3 then shows the process of realizing card-friend credit with the smart loan review system. This process embodiment describes an embodiment of the process in which customers of the financial institution themselves conduct card-friend credit.

At the beginning, the loan applicant filed a loan application with a financial institution through the card friend credit application page 301. Since the loan applicant may be a bank customer, if it has already saved the procedures of identity authentication and blacklist checking, you can directly register with the card friend Select option 303 on the credit application page 301, and the system will execute the issuance check joint 305, risk check 307, and model quotation 309.

In the step of issuing the joint collection 305, the credit scoring model is used to obtain the results of the collection review based on the applicant's basic information, transaction information, and digital browsing behavior. In the step of risk checking 307, blacklist screening is performed using a blacklist screening model. In the model quotation 309, the undertaking value model is used to generate a pricing table for the loan applicant through the undertaking value model based on the basic information of the loan applicant and various records of transactions with financial institutions. After that, after sending the insurance newsletter 311 to execute the online insurance 313, and finally the loan application is approved, 315 is allocated.

For the method flow implemented by the smart loan review system, please refer to the flowchart of the embodiment of the smart loan review method shown in FIG. 4. According to the embodiment, the method runs in the main server in the embodiment shown in FIG. , The main server is like a computer system, which is equipped with a processor and memory to execute a software-implemented smart loan review method.

At the beginning of the process, in step S401, the user can submit a loan application request through the user interface provided by the system. This is the system requiring the loan applicant to provide basic information and relevant loan application information. In step S403, the system checks according to the user's identity, and checks the user-related historical data, such as the user's credit information, transaction data, etc., from the database 41 (such as a database in a financial institution). In step S405, the identity of the user can be determined. For example, the past data of the loan applicant can be obtained based on the data of the financial institution, and it can be judged to be a good customer (existing customers of a specific financial institution, which can directly obtain financial proof), internal Blacklist, or no record, etc. After that, in step S407, the money laundering list is queried, and the smart loan review system will perform the money laundering black name through the blacklist screening model 42 Single screening, the scope of the blacklist query includes Internet addresses (IP), blacklists in financial institutions, and blacklists provided by the Money Laundering Prevention Center. The blacklist screening model 42 can be composed of the Chinese and English names and names of the applicant Eliminate errors in data such as changes and related party information, and get correct screening results.

Furthermore, the smart loan review system uses machine learning algorithms to train based on the internal data of financial institutions and the blacklists of external money laundering prevention centers to establish a blacklist screening model 42, which can effectively screen loan applicants. What is running is that when the system obtains the domestic and foreign money laundering list (blacklist), it can build a screening list model through machine learning. The learned data also includes the data that has used the accepted financial institutions to apply and passed the loan application, and records Good or bad data, related data through continuous learning, including various data fed back to the system, can optimize the list screening model and reduce false judgments. In particular, the customer groups that can be quickly applied for loans are set by financial institutions. When the applicants meet a specific customer group, the efficiency of list screening can be increased.

In step S409, the system measures the credit risk of the loan applicant, and can obtain the credit information of the loan applicant from the joint credit investigation center and the customer credit information obtained by the internal examination system of the financial institution, and enter the real-time joint application information and basic information , Transaction data, digital browsing data and other data are sent to the credit scoring model 43 to obtain the credit investigation results. In step S411, the system inputs the cost parameters, basic data, and transaction data of various financial products of the financial institution into the undertaking value model 44, and executes undertaking value judgments, so that the system can make decisions based on the comprehensive consideration of the applicant’s risk and value To issue a pricing table, such as step S413, to generate loan application data. In step S415, each communication method is used to guarantee the loan applicant. The guarantee method is, for example, a text message (or a specific communication method) to confirm the loan details and the loan applicant. Condition, after the guarantee is completed, in step S417, funds are allocated based on the account information provided by the loan applicant.

For example, when the system receives a loan request from a loan applicant through the web interface proposed by the main system, it is transmitted to the intelligent service platform in the main server, and the blacklist screening model 42 in the system obtains the screening results. Model 43 obtains the result of the joint collection, and then the commitment value model 44 obtains the loan plan. After integrating the results of each intelligent model, it is based on the pre-planned risk pricing The table generates loan plans, including quotas, interest rates, and repayment plans.

Fig. 5 shows a flowchart of an embodiment of smart verification in a smart loan verification method.

In this process, the system obtains internal verification data based on the loan applicant's data (step S501), and connects to the database of international anti-money laundering agencies such as the Money Laundering Prevention Center (step S503) to cross-check the internal data (step S505) ), and the intelligent model is introduced at this time (step S507). The intelligent model is the blacklist screening model described in the above embodiment, except that the internal data of financial institutions and the data of various lenders' loan applications are used for training with machine learning algorithms ( Step S509) The blacklist screening model is established. On the one hand, the blacklist prediction is performed in real time (step S511). After obtaining the loan applicant’s information, the blacklist screening model is used to determine whether the loan applicant is a money laundering prevention risk account ( Step S513).

According to practical examples, the blacklist screening model obtained through training and learning can effectively increase the screening accuracy (precision). When the model predicts to be true and the actual is also true, the screening accuracy is, for example, when 50 people are The intelligent model predicts that it is a specific blacklist, but it actually confirms that there are 25 people on the blacklist, and the accuracy rate is 50%. There is also a recall rate (recall). When the blacklisted applicants are predicted by the smart model, the ratio is true. For example, a total of 10 people are actually on a specific blacklist, and the smart model predicts that 2 of them are For specific blacklists, the recall rate is 20%.

In one embodiment, if it is determined that the loan applicant is not on the blacklist, the follow-up process of the entire smart loan review method can be continued; on the contrary, if the loan applicant is determined to be on the blacklist, it may be converted to manual judgment at this time. List, that is, reject loan applications, and an alert condition can be designed for this.

Fig. 6 then shows a flowchart of an embodiment of smart quotation in the smart loan review method.

The smart quotation process shown first obtains the credit information of the loan applicant (step S601), which is the internal information of a specific financial institution, such as customer basic information, credit card transactions, recurring information, channel response information, and customer entry information, etc. , The process continues to obtain the external information of the loan applicant (Step S603), such as browsing information on the official website of a financial institution, off-site publicity information, social media information, etc.

These internal and external data are introduced into a credit scoring model (step S605), credit scoring is performed, and risk pricing is further determined (step S607), and then a commitment value model is introduced (step S609) to determine a loan plan, including setting a loan limit, Interest rate, time, repayment plan, setting default clauses, etc. (step S611).

To sum up, according to the smart loan review system described in the above embodiments, an unmanned online loan service process is proposed, and machine learning methods are specially introduced to establish smart models, such as money laundering prevention models, credit scoring models, and contractors. The value model integrates customer information to comprehensively evaluate risks, and provides quotes for quotas, interest rates, and futures. The intelligent model can accurately measure the risk and value of the loan applicant, and provide an appropriate credit limit to improve risk management capabilities, pending confirmation by the loan applicant Funds are allocated in time after the loan conditions are correct, to provide users with comprehensive, fast and convenient personal credit loan services, and to generate a new innovative service structure through series integration of processes under compliance with laws and regulations. No one is involved in the application, review, guarantee, and funding stages , Its main goal is to quickly complete the application and loan approval.

The content disclosed above is only a preferred and feasible embodiment of the present model, and does not therefore limit the scope of the patent application of the present model. Therefore, all equivalent technical changes made using the description and schematic content of the present model are included in the application of the present model. Within the scope of the patent.

10: Internet

100: main server

105: Intelligent Service Platform

151: Smart Check Module

152: Credit Scoring Module

153: Valuation Module

101: Web server

103: Database

107: Examination system

108: Protection System

109: Accounting System

110: Risk Database

14: Web interface

15: User

16: Money Laundering Prevention Center

18: Joint Credit Information Center

Claims (10)

An intelligent loan review system, including: a main server, which runs an intelligent service platform, is equipped with an intelligent verification module, a credit scoring module, and a value evaluation module, which are learned and trained by machine learning methods. A blacklist screening model, a credit scoring model, and an undertaking value model, wherein the main server provides a user interface through the Internet for users to apply for loans; a risk database on the smart service platform When the verification module establishes the blacklist screening model, it connects to the risk database and an external money laundering prevention center to obtain a blacklist inside and outside a financial institution; a collection review system to record the joint collection data of the loan applicant , Basic information of the loan applicant, historical transaction data and digital browsing data; a pair of guarantee system, used to execute the guarantee with the loan applicant; an accounting system, used to execute the funding after the completion of the guarantee; the main server The intelligent loan review of the operation in China includes: a loan applicant submits a loan application request to the financial institution; in the financial institution, after identification, the blacklist screening model is used to determine whether the loan applicant is a money laundering blacklist ;According to the identity of the loan applicant, obtain the credit information of the loan applicant from the financial institution, a joint application information, the basic information of the loan applicant, transaction data and digital browsing data, and enter a credit scoring model to obtain A credit investigation result; input the loan application data of the lender, the cost parameters of the financial institution, basic information, and the transaction data of various financial products to an undertaking value model, and perform an undertaking value judgment, and make a reservation for the loan applicant Formulate a pricing table to determine a loan plan; and in the guarantee system, after guaranteeing the loan applicant, the accounting system will allocate funds. The smart loan review system according to claim 1, wherein the blacklist screening model in the main server is established by machine learning based on the internal data of the financial institution and the blacklist of the money laundering prevention center Intelligent model. Such as the smart loan review system described in claim 2, wherein the basic information, transaction information, joint solicitation information and digital browsing behavior of the loan applicant obtained by the financial institution are input into the blacklist screening model to filter the money laundering blacklist . The smart loan review system according to claim 1, wherein the credit scoring model in the main server is based on the credit data, transaction data, personal historical data of the financial institution and a joint credit center The intelligent model established by the obtained joint solicitation data. The smart loan review system according to claim 4, wherein the transaction data input to the credit scoring model includes credit card consumption, counter deposit foreign exchange transactions, and various financial product transactions generated by the loan applicant at the financial institution. The smart loan review system according to claim 1, wherein the commitment value model in the main server is data that uses a machine learning method to evaluate the commitment value according to the knowledge provided by the financial institution, the historical data of loan approval, and The result, and the intelligent model established by the financial institution’s cost and risk information. The smart loan review system according to claim 6, wherein the financial institution continuously provides new cost data and loan interest rates to update the commitment value model. The smart loan review system according to any one of claims 1 to 7, wherein when guaranteeing the loan applicant, confirm the loan details and conditions with the loan applicant through a communication method, and the guarantee is completed. The smart loan review system according to any one of claims 1 to 7, wherein the loan plan includes setting a loan limit, interest rate, time, repayment plan, and a default clause. The smart loan review system according to any one of claims 1 to 7, wherein The main server is equipped with a web server, whereby the main server provides the user interface through the Internet for users to apply for loans, and uses the user interface to allow users to fill in the fields required for loan applications Information, including personal data, loan application requirements, and financial proof.

Images