DESIGN AND IMPLEMENTATION OF A COMPUTERIZED LOAN MANAGEMENT SYSTEM FOR REJECTING OR APPROVING LOAN REQUEST USING CREDIT RISK AND EVALUATION MODELS
Abstract : Nowadays, home loan is a frequently accessed component of people’s financing activities. Homeowners wants to increase the probability of loan acceptance, however banks seek to borrow money to low risk customers.
This paper compared and examined the machine learning models to select when loan applicants evaluating their probability of success. This paper introduced the recommended models for the problem, explanations on how to use the selected model.
6 candidate models, including Logistic regression, Decision tree, Random Forest, support vector machine (SVM), Ada Boost and Neural Network are selected. The model selection process would focus on the model’s accuracy on test data as well as the interpretability of these models.
The models’ result was interpreted to derive optimal strategies to be undertaken by both debtors and creditors. Throughout comparison between these models, logistic regression was the best in terms of interpretability and accuracy.
Nonetheless, other models could bolster the decision-making process by examining their confusion matrices and the fitted importance of predictors in each model. This paper revealed practical implications of machine learning theories on home loan approval and credit risk and aimed to help decision making for both debtors and creditors.
On the other hand, the prediction model helps us to select the most convenient decision by means of repeated behavior patterns [8]. Similarly, these models are key pieces to be able to understand the behavior and create strategies that can cause greater impact [9].
Therefore, in the "Predictive model of delinquency rank for bank loans using simulated data", the behavior of customers was determined by means of a statistical model in order to simulate data and predict the delinquency of a bank loan [10].
Considering the above, the use of neural networks is easily adjusted to financial drawbacks. One of the real cases is to apply it in calculating the probability of default in the portfolio of the financial company [11].
Moreover, to define the best method to evaluate credit risk, Binary Regression has been proposed in conjunction with Artificial Neural Networks using the mathematical learning algorithm Multilayer Perceptron [12].
Likewise, [13] comments that improving the neural network by GA algorithm avoids the limitation of the evaluation process, making the alert result more reliable and objective. [14] reports that it is necessary to improve credit risk models for the sector to progress.
Complexity and Cost
Development Complexity: Creating a sophisticated LMS that integrates credit risk models involves complex software development and system integration. This can require significant technical expertise and time.
High Costs: Developing, deploying, and maintaining such a system can be expensive. Costs include software development, hardware infrastructure, ongoing maintenance, and potential updates to models.
Data Privacy and Security
Data Sensitivity: Handling personal and financial data of loan applicants requires stringent security measures. Breaches can lead to legal issues and damage to the institution's reputation.
Compliance Issues: The system must comply with regulations like GDPR or CCPA, which can be challenging to manage and may require continuous updates.
Model Limitations
Risk of Bias: Credit risk models can inadvertently incorporate biases if not carefully designed and tested. This can lead to unfair discrimination against certain groups of applicants.
Model Accuracy: Models are only as good as the data they are trained on. Inaccurate or incomplete data can lead to poor decision-making and increased risk for the lender.
System Reliability
Technical Failures: Dependence on computerized systems means that technical failures, such as software bugs or hardware malfunctions, can disrupt loan processing and decision-making.
Maintenance: Ongoing system maintenance and updates are necessary to ensure reliability and incorporate new features or changes in regulations.
The computerized Loan Management System (LMS) is designed to enhance the efficiency and accuracy of the loan approval process through the integration of advanced credit risk and evaluation models.
This system is structured around several key modules, each addressing crucial aspects of loan management. The User Management Module ensures secure access and role-based permissions for applicants, loan officers, and administrators.
The Application Management Module facilitates the online submission and tracking of loan applications, streamlining the initial interaction with the system. The Credit Risk Assessment Module employs sophisticated credit scoring algorithms and risk analysis techniques to evaluate applicants' creditworthiness, ensuring that decisions are data-driven and objective.
The Evaluation and Decision Module automates the approval or rejection process based on predefined criteria, while also providing tools for manual review when needed. This module ensures that every application is processed efficiently and transparently.
Increased Efficiency
Faster Processing: Automated systems can process loan applications much faster than manual methods, reducing the time required to approve or reject applications.
Streamlined Workflow: Automation reduces the need for repetitive tasks, allowing loan officers to focus on more complex aspects of loan management.
Improved Accuracy
Consistent Decision-Making: Automated systems apply the same criteria to all applications, reducing the risk of human error and ensuring consistent decision-making.
Advanced Analytics: Credit risk models can analyze vast amounts of data to provide accurate risk assessments and creditworthiness evaluations.
|