JP7391497B2 - Loan screening device - Google Patents

Description

TECHNICAL FIELD The present invention relates generally to financial technology (fintech), and more particularly to loan screening systems.

Currently, financial institutions provide various financial services not only through their branches but also through networks such as websites. For example, when applying for a personal loan such as a home loan, a customer accesses the financial institution's website, enters the necessary information in each input field of the examination application form in a prescribed format, and provides information such as a driver's license. An application for loan examination can be made by transmitting the image information read from predetermined identity verification materials.

Japanese Patent Application Publication No. 2017-146844

When applying for a home loan from such websites, a preliminary screening or provisional screening service is provided, and it is determined whether the desired loan amount is within the borrowable amount based on information such as annual income that the customer entered in the screening application form. Customers can understand the extent of preferential interest rates that can be provided. However, the examination content is based on information self-reported by the customer, and we check whether the entered information is accurate, whether there are any problems with the customer's personal credit information, and whether the customer is related to anti-social forces (anti-social forces). Financial institutions are unable to sufficiently confirm the following:

In a typical loan application system, as shown in Figure 1, there are various channels such as web application, tablet application, and in-store application, and with web application and tablet application, customers can easily apply for preliminary screening. can. In the bank examination after application, the person in charge checks the entered information based on the submitted identity verification documents, and uses a points system to examine the information based on age, place of work, years of service, annual income, desired loan amount, etc. We manually check personal credit information, such as whether loan repayments are in arrears, and whether there is any connection to anti-social forces. The examination results are then notified to the customer after being confirmed by other approvers. For this reason, the screening process takes time, and the results of the screening process may take several days to complete, so there is a risk that customers may leave the company before signing a contract. Furthermore, since the work is done manually, the administrative burden on the person in charge is heavy.

In view of the above-mentioned problems, an object of the present invention is to provide a technique for realizing an efficient loan screening system.

In order to solve the above problems, one aspect of the present invention includes a reception unit that receives input examination application information and identity verification materials, and executes a loan examination for a user based on the examination application information and the identity verification materials. The present invention relates to a loan examination device having an examination processing section.
Regarding.

According to the present invention, an efficient loan examination system can be realized by automating not only the reception of loan examination applications but also the loan examination.

It is a schematic diagram showing conventional loan examination processing. 1 is a schematic diagram showing a loan screening system according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing loan screening processing according to an embodiment of the present invention. FIG. 1 is a block diagram showing the hardware configuration of a loan screening device according to an embodiment of the present invention. FIG. 1 is a block diagram showing the functional configuration of a loan screening device according to an embodiment of the present invention. It is a flow chart showing loan screening processing according to an embodiment of the present invention.

Embodiments of the present invention will be described below based on the drawings.

In the following embodiments, a loan screening device for screening loans applied via a network will be disclosed. To summarize an embodiment described later, a loan screening device executes a loan screening for a user based on screening application information and identity verification materials received from the user via a network. More specifically, the loan screening device determines whether the screening application information is valid, that is, whether the input contents of the screening application information and the contents of the personal identification materials match, and if they match, The learned loan screening model is used to determine whether or not a loan is available based on the screening application information.

Although the following embodiments will mainly be explained with respect to housing loans, the loans according to the present invention are not limited thereto, and can be applied to various personal loans such as card loans, education loans, private car loans, and various business loans by corporations. may be applied.

First, with reference to FIG. 2, a loan screening system according to an embodiment of the present invention will be described. FIG. 2 is a schematic diagram showing a loan screening system according to an embodiment of the present invention.

As shown in FIG. 2, the loan screening system 10 includes a user terminal 50, a loan screening device 100, an in-house database (DB) 150, and an external inquiry device 200.

The user terminal 50 may be any information processing device with a communication function, such as a personal computer, smartphone, or tablet, and accesses the website of the financial institution and executes a loan screening application via the network. do.

As will be explained in detail below, when the loan screening device 100 receives a loan screening application from the user terminal 50, it executes the loan screening for the applied loan by communicating with the bank's in-house DB 150 and the external inquiry device 200. .

For example, in the specific example shown in FIG. 3, when the loan screening device 100 receives screening application information and identity verification materials for preliminary screening from the user terminal 50 via a website operated by the financial institution, the loan screening device 100 The validity of the examination application information is determined by specifying the written contents from the image information of the confirmation materials and determining whether the input contents of the examination application information match the specified written contents. When determining that the screening application information is appropriate, the loan screening device 100 uses a trained artificial intelligence (AI) model based on the screening application information and, if available, the customer's transaction history information. Perform a pre-screening. After notifying the user terminal 50 of the preliminary screening results and accepting the main screening application, the loan screening device 100 receives the main screening application information received from the customer (typically, in addition to the information for the preliminary screening, collateral (including information regarding the property, etc.), the final examination result is notified to the user terminal 50 based on the external inquiry results from the external inquiry device 200, etc.

Although the illustrated example describes a loan screening device 100 that follows a two-step screening process of preliminary screening and main screening that is often performed for home loans, the loan screening device 100 according to the present invention is applicable to this process. Without limitation, for example, for unsecured loans such as card loans, the approval or rejection of the loan is determined through a one-stage examination that includes both an internal loan determination using screening application information and internal bank information, and an external inquiry determination. It's okay.

The in-house DB 150 stores various information held by the financial institution, such as customer information and transaction history information. For example, when the loan screening device 100 requests information regarding a user who has applied for a loan screening, the in-house DB 150 provides the loan screening device 100 with transaction history information such as deposit/withdrawal information and loan information of the user's account. .

The external inquiry device 200 is typically an inquiry system of a loan guarantee company or an inquiry system of the Deposit Insurance Corporation. Provides information, anti-social check results, etc.

The loan screening device 100 is typically realized by a server, and may have a hardware configuration as shown in FIG. 4, for example. That is, the loan examination device 100 includes a drive device 101, an auxiliary storage device 102, a memory device 103, a CPU (Central Processing Unit) 104, an interface device 105, and a communication device 106 that are interconnected via a bus B.

Various computer programs including programs for realizing various functions and processes to be described later in the loan examination device 100 may be provided by a recording medium 107 such as a CD-ROM (Compact Disk-Read Only Memory). When the recording medium 107 storing the program is set in the drive device 101, the program is installed from the recording medium 107 to the auxiliary storage device 102 via the drive device 101. However, the program does not necessarily need to be installed on the recording medium 107, and may be downloaded from any external device via a network or the like. The auxiliary storage device 102 stores installed programs as well as necessary files and data. The memory device 103 reads programs and data from the auxiliary storage device 102 and stores them when there is an instruction to start a program. The CPU 104, which functions as a processor, executes various functions and processes of the loan examination device 100, which will be described later, in accordance with various data such as programs stored in the memory device 103 and parameters necessary to execute the programs. The interface device 105 is used as a communication interface for connecting to a network or external device. The communication device 106 executes various communication processes for communicating with external devices. However, the loan screening device 100 is not limited to the hardware configuration described above, and may be realized by any other suitable hardware configuration.

Next, referring to FIG. 5, a loan examination device 100 according to an embodiment of the present invention will be described. FIG. 5 is a block diagram showing the functional configuration of a loan screening device according to an embodiment of the present invention.

As shown in FIG. 5, the loan screening device 100 includes a reception section 110 and a screening processing section 110.

The reception unit 110 receives the input screening application information and identity verification materials. More specifically, the reception unit 110 determines whether the examination application information is valid based on the examination application information input by the user and the identification information provided by the user. Specifically, the reception unit 110 receives examination application information and image information of identity verification materials (for example, PDF data, JPEG data, PNG data, GIF data, etc.) received from the user terminal 50 via the network, and performs the examination. The validity of the review application information is determined by determining whether the input content entered in the application information matches the content written in the specified identity verification material.

For example, the screening application information is provided on the website of the financial institution and has a predetermined format, and for example, the customer is asked to enter information into blank fields in the format. On the other hand, the identity verification material is a predetermined type of certificate issued by a public institution, such as a driver's license or my number card, and layout information of the certificate is registered for each type. Typically, the layout information may indicate in which area within the image of each certificate information such as a name, address, date of birth, etc. is written.

The reception unit 110 analyzes or recognizes the image information of the identity verification material, determines the type of the submitted identity verification material, and updates each entry in the identity verification material based on the layout information of the registered certificate of the type. The content can be specified. Note that when a certificate written in an unregistered layout is submitted, the reception unit 110 asks the user terminal 50 where the written content necessary for determining the validity of the examination application information is written. You may also inquire and have them specify the location where it is written.

The examination processing unit 120 executes a loan examination for the user based on the examination application information and identification materials. More specifically, when the examination processing unit 120 determines that the examination application information is valid based on the identity verification materials, the examination processing unit 120 executes a loan examination for the user based on the examination application information using the learned model. Specifically, when the reception unit 110 determines that the screening application information provided from the user terminal 50 is valid based on the identification materials, the screening processing unit 120 performs loan screening based on the screening application information. The system determines whether the applied loan can be financed or not, and whether the entire or part of the desired loan amount can be financed.

The loan availability determination can be realized using an artificial intelligence (AI) model such as a neural network or a machine learning model that is trained in advance based on examination cases held by the financial institution. For example, a model for mortgage screening may be learned by any suitable learning algorithm using training data sampled from past mortgage loan cases made by financial institutions. For example, without limitation, if the model is configured by a neural network, information regarding each past mortgage loan case, such as borrower attribute information (age, area of residence, period of residence, rental, home ownership, Residence type (condominium, house, etc.), work type (self-employed, company employee, contract employee, etc., workplace, years of service, etc.), income/asset information (annual income, existing loans, financial assets, real estate, etc.), loan condition information (housing) Purchase price, down payment, desired loan amount, repayment period, etc.), collateral information (address, size of the house to be purchased, etc.) are prepared as learning input data, and repayment results corresponding to each input data (completed payment, (Unpaid, etc.) may be prepared as learning output data, and the neural network may be trained by backpropagation or the like using the prepared learning data until the prediction error becomes equal to or less than a predetermined threshold. Alternatively, the repayment result is not limited to the two outputs of fully paid and unpaid; it is fully paid according to the original repayment plan, the repayment plan has been revised but the repayment is completed, the repayment is recovered due to a guarantee from a guarantee company, and the repayment is not completed and is recovered by auction etc. , more detailed output such as unpaid and unrecoverable. Alternatively, a rating or the like may be output, and the screening processing unit 120 presents customers with a high rating with preferential interest rates corresponding to the rating, and provides customers with a low rating with a loan that is not available or a reduction in the desired borrowing amount. It is also possible to present a judgment on whether the loan is possible with conditions. Those skilled in the art will easily understand that machine learning models compatible with these various output formats can be learned by appropriately adjusting the learning data.

The examination processing unit 120 notifies the user terminal 50 of the examination result. For example, the screening result may be that a loan is possible or impossible based on all loan conditions such as the user's desired borrowing amount and repayment period, or it may be possible to provide a loan based on all loan conditions such as the user's desired loan amount and repayment period, or it may be possible to provide a loan based on all loan conditions such as the user's desired loan amount and repayment period. It may be possible to make a loan based on the determination that the loan is possible.

In one embodiment, the screening processing unit 120 obtains transaction history information and performs loan screening. More specifically, the screening processing unit 120 may acquire the user's transaction history information, and perform loan screening for the user based on the screening application information and transaction history information using a learned model. For example, if the user already has a transaction with the financial institution, loan screening may be performed using transaction history information of the user in addition to screening application information. In this case, the trained model is trained by including the borrower's past transaction history information (deposit/withdrawal history, borrowing history, etc.) in the training input data in addition to the various information regarding each past mortgage loan case described above. may be done. This will incorporate time-series information such as the borrower's past income and expenditure patterns, trends in income and expenditure balance, and stability of income into the learning data, making it possible to construct a more accurate loan appraisal model.

Furthermore, in one embodiment, the examination processing unit 120 may further inquire about the user to the external inquiry device 200, and determine whether to perform a loan examination for the user based on the inquiry result. In other words, the examination processing unit 120 can perform loan examinations by taking into account not only the bank's internal loan approval determination, but also personal credit information from the loan guarantee company, antisocial check results from the Deposit Insurance Corporation, and the like. This makes it possible to perform more appropriate loan examinations based not only on self-reported information from users but also on various types of information from external inquiry institutions.

Next, referring to FIG. 6, a loan examination process according to an embodiment of the present invention will be described. FIG. 6 is a flowchart showing loan screening processing according to an embodiment of the present invention. The loan screening process may be started, for example, in response to reception of screening application information from the user terminal 50.

As shown in FIG. 6, in step S101, the loan screening device 100 receives screening application information from the user terminal 50 via the network.

In step S102, the loan screening device 100 receives identity verification materials from the user terminal 50 via the network. Note that steps S101 and S102 may be performed in different orders or may be performed together.

In step S103, the loan screening device 100 determines the validity of the screening application information based on the identification data. If the input contents of the examination application information and the contents written in the identification document match, the loan examination device 100 determines that the examination application information is valid, and proceeds to step S104. On the other hand, if the input content of the screening application information and the content described in the identification document do not match, the loan screening device 100 notifies the user terminal 50 of the discrepancy and prompts the user to check whether there are any errors in the input content. It's okay.

In step S104, the loan screening device 100 executes a loan screening for the user based on the screening application information using the learned model. After the loan examination, the loan examination device 100 notifies the user terminal 50 of the examination result. For example, the screening result may be that a loan is possible or impossible based on all loan conditions such as the user's desired loan amount and repayment period, or it may be possible or impossible to make a loan based on a review of some loan conditions. It may be that.

Although the embodiments of the present invention have been described in detail above, the present invention is not limited to the specific embodiments described above, and various modifications can be made within the scope of the gist of the present invention described in the claims. - Can be changed.

10 Loan screening system 50 User terminal 100 Loan screening device 200 External inquiry device

Claims (3)

a reception unit that accepts the input screening application information and identity verification materials , and identifies the contents of the provided identity verification materials based on the arrangement state of the information present in the image of the identity verification materials;
If the screening application information and the identity verification materials match, the information will be output when the user's desired loan condition information, attribute information, income/asset information, and collateral information included in the screening application information are input into a machine learning model. Based on the rating, if the rating is high, a preferential interest rate corresponding to the rating is determined; if the rating is low, it is determined whether the loan is not available under the desired loan terms or whether the loan amount is reduced. A loan examination device having an examination processing unit that performs The loan screening device according to claim 1, wherein the screening processing unit acquires transaction history information and uses the machine learning model to determine whether or not a loan can be granted to the user based on the screening application information and the transaction history information. The screening processing unit further refers the user to an external inquiry system, and determines whether or not the user can receive a loan based not only on the loan availability determination within the financial institution but also on the personal credit information and antisocial check results of the external institution. , The loan examination device according to claim 1 or 2.

Images