KR-20260062119-A - AI algorithm-based loan risk management method

KR20260062119AKR 20260062119 AKR20260062119 AKR 20260062119AKR-20260062119-A

Abstract

The present invention relates to a loan risk management method based on an artificial intelligence algorithm, characterized by comprising: a first step of collecting data necessary for managing insurance collateral loan risk for an insured customer; and a second step of estimating data for insurance collateral loan risk management using the collected data based on an artificial intelligence (AI) model to predict whether the data exceeds or falls short of a threshold. According to the present invention, data necessary for risk management when providing secured loans to customers who have subscribed to various insurances, such as life insurance, is collected and analyzed based on artificial intelligence (AI) learning models such as LSTM and GAN to estimate necessary data and predict whether it exceeds or falls short of a threshold, thereby enabling stable insurance loan services to customers, as well as effectively managing loan risks by managing the decline in customer creditworthiness and crisis situations.

Inventors

김남도
정재훈
임광수

Assignees

주식회사 시안소프트

Dates

Publication Date: 20260507
Application Date: 20241025

Claims (5)

Step 1: Collecting data necessary for managing insurance collateral loan risks for insured customers; and A second step of estimating data for insurance collateral loan risk management using the collected data based on an artificial intelligence (AI) model to predict whether the value exceeds or falls short of a threshold; An artificial intelligence algorithm-based loan risk management method characterized by including
In Article 1, The above second step acquires individual factors and unique risks through an artificial intelligence (AI) model, and The personal factors obtained above include "predicted credit scores" and "predicted DSR ratios" as data for scale analysis based on the customer's economic activities, and An artificial intelligence algorithm-based loan risk management method characterized by the above-mentioned intrinsic risk including "insurance premium payment diligence index (index: maximum score for diligent payment, deducted for non-payment, delinquency, etc.) - insurance payment diligence probability (Risk)" and "loan payment diligence index (index - maximum score for diligent loan repayment, deducted for non-payment, delinquency, etc.) - loan repayment diligence probability (Risk)" as the probability of maintaining insurance and insurance-backed loans.
In Article 1, The data collected in the first stage above includes personal factors, which are data for scale analysis based on the customer's economic activity; unique factors regarding the customer's insurance and insurance-backed loans; and environmental factors for identifying risk factors by analyzing the environment affecting the customer's economic activity. An artificial intelligence algorithm-based loan risk management method characterized in that the second step above is a step of predicting insurance collateral loan risk by learning collected data using an artificial intelligence (AI) model called a GAN.
In Paragraph 3, The above second step involves constructing a hexahedron with individual factors as the A-axis, intrinsic factors as the B-axis, and environmental factors as the C-axis to form a single cell, thereby creating a time domain; An artificial intelligence algorithm-based loan risk management method characterized by learning the intrinsic factors and environmental factors of the above cell as time series data using LSTM (Long Short-Term Memory), and learning the intrinsic factors using GAN (Generative Adversarial Network).
In Paragraph 3, The above second step involves constructing a hexahedron with individual factors as the A-axis, intrinsic factors as the B-axis, and environmental factors as the C-axis to form a single cell, thereby creating a time domain; An artificial intelligence algorithm-based loan risk management method characterized in that the intrinsic factor and environmental factor of the above cell are time-series data, are trained using LSTM (Long Short-Term Memory), and the intrinsic factor is trained using any one of cGAN (Conditional GAN), DCGAN (Deep Convolutional GAN), or LSGAN (Least Squares GAN).

Description

AI algorithm-based loan risk management method The present invention relates to a loan risk management method based on an artificial intelligence algorithm, and more specifically, to a method for managing loan risk through the management of creditworthiness decline and crisis by supporting the stable use of insurance loan services by customers based on an artificial intelligence (AI) algorithm. Generally, life insurance is an insurance product that pays out a benefit when the insured dies or a specific life-related event occurs. It is primarily purchased to financially protect family or loved ones, and there are various types available. Loan products offered to customers who have subscribed to such life insurance are broadly classified into insurance payout-backed loans and insurance contract-based credit loans. In this case, the insurance benefit secured loan determines the loan amount and interest rate by evaluating the customer's creditworthiness based on the premiums paid, while the aforementioned insurance contract-based credit loan determines the loan amount and interest rate by evaluating the premiums and creditworthiness of customers who have contracted insurance and paid premiums exceeding a certain percentage set by the insurance company. Meanwhile, as the macroeconomic environment significantly impacts the risk management of life insurance loan products, continuous monitoring of various economic indicators and customers, as well as appropriate responses, are necessary. As prior art regarding loan product services for such insurance customers, the following have been proposed: Published Patent No. 10-2002-0035639 (Reference 1), ‘a financial loan service system on the Internet using insurance products’; Published Patent No. 10-2003-0054225 (Reference 2), ‘a savings/loan system and method using product purchase interest rates’; Registered Patent No. 10-0791650 (Reference 3), ‘a financial consulting system for customer risk situations, a method and a computer-readable recording medium recording a program for executing the same’; and Published Patent No. 10-2001-0082393 (Reference 4), ‘a financial loan service method via the Internet’. However, the conventional insurance-based loan services proposed through the aforementioned references do not provide customized risk management for current customers (such as insurance cancellation or loss of the benefit of time due to non-payment of loan installments), and in particular, since insurance-based loans use paid insurance premiums as collateral, they overlook the need for risk management. FIG. 1 is an overall system configuration diagram for loan risk management based on an artificial intelligence algorithm according to the present invention. FIG. 2 is a diagram illustrating the materials (data) required for loan risk management based on an artificial intelligence algorithm according to the present invention. FIG. 3 is a diagram illustrating the cell configuration for analyzing the influence of personal factors, unique factors, and environmental factors required for loan risk management based on an artificial intelligence algorithm according to the present invention. FIG. 4 is a diagram illustrating the cell configuration for analyzing the influence of personal factors, unique factors, and environmental factors required for loan risk management based on an artificial intelligence algorithm according to the present invention. Figure 5 is an LSTM application diagram for loan risk management based on an artificial intelligence algorithm according to the present invention. Figure 6 is a diagram illustrating the B (inherent factor) and C (environmental factor) domains derived through the LSTM model (algorithm) of Figure 5. Figure 7 is a conceptual diagram of a GAN applied for loan risk management based on an artificial intelligence algorithm according to the present invention. FIG. 8 is a diagram illustrating the data input/output relationship for the present invention from the GAN concept diagram of FIG. 7. The features of the artificial intelligence algorithm-based loan risk management method according to the present invention can be understood through embodiments described in detail below with reference to the attached drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention. Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely the most preferred embodiments of the present invention and do not represent all of the technical ideas of the present invention; thus, it should be understood that various equivalents and modifications that can replace them may exist at t