LOAN APPROVAL PREDICTION USING

MACHINE LEARNING
Abstract:
The payment for acquiring assets keeps on increasing with each passing day and the capital
required to purchase an asset in entirety is extremely high. So it is not possible to purchase it
out of your savings. Loans are a good way of getting the required funds. However, taking out
a loan is a lengthy process. There are so many steps that need to be completed and there is
also the possibility that it might not get approved. To reduce the time required to approve the
loan and the risk that comes along with it, many loan prediction models were devised. The
aim of this project was to study as well as compare different prediction models for loan
approval and denial to find out which one has the smallest margin of error while predicting
whether or not the loan should be approved using a risk analysis. In the course of the
analysis, it was determined that the most accurate prediction model is the one which is based
on Random Forest. This will help in minimizing the time and human effort needed to process
loans and identify the best prospects for funding.
Keywords: Loan, Machine Learning, Training, Testing, Prediction.
Introduction:
In banking and financial institutions, approval of loans is essential for institutional growth
and stability. Traditional evaluation of loan proposals is dependent on manual validation and
risk-based checks, which take time and can be subject to errors, thus resulting in over-
approval or under-approval of undeserving or default-prone clients, respectively. For this
problem, machine learning (ML) models can eliminate delays and augment predictive
accuracy regarding the approval of loans, allowing improved decision making, minimizing
default in loans, and maximizing asset deployment.
Simple ML algorithms such as Logistic Regression, Decision Trees, and Support Vector
Machines (SVMs) have been applied to loan prediction systems with satisfactory outcomes.
But when loan data is increasingly complex, there is a requirement for more complex models
to handle complex relationships and enhance prediction accuracy. Class imbalance, where
approved loans outnumber defaults by far, is one of the major challenges in loan prediction.
This may skew the predictions towards the majority class (approved loans). To address this,
sophisticated models like Random Forests, XGBoost, Neural Networks, and SVM have been
used to enhance accuracy and manage class imbalanceThis paper examines the use of these
machine learning models to predict loan approval improvement, based on features such as
credit score, income, amount of loan, and employment status. The decision-making process is
automated by the system, aiding financial institutions to reduce risks and make the approval
process more efficient, while being free from prejudice and making informed decisions.
The subsequent sections deal with the training dataset, the algorithms used, their evaluation
parameters, and results achieved.
Related Work
Several studies have explored the use of machine learning techniques for loan approval
prediction to improve decision-making efficiency and accuracy. Traditional methods like
Logistic Regression (LR) and Linear Discriminant Analysis (LDA) were initially applied, but
they faced limitations in handling complex relationships within the data. Recent
advancements have favored Decision Trees (DT), Naive Bayes, and Support Vector Machines
(SVM), achieving moderate success in loan prediction tasks (Menzies, Greenwald, & Frank,
2007; Okutan et al., 2014).
In terms of performance, ensemble methods like Random Forests (RF) and Gradient Boosting
Machines (GBM), including XGBoost and LightGBM, have outperformed traditional models
due to their ability to handle large datasets with complex feature interactions (Lessmann,
Baesens, Mues, & Pietsch, 2008). Additionally, Neural Networks (NN), particularly
Feedforward Neural Networks, have demonstrated good predictive power by modeling non-
linear relationships within loan data.
A key challenge in loan approval prediction is the class imbalance problem, as approved
loans vastly outnumber defaults. Methods like SMOTE (Synthetic Minority Over-sampling
Technique) and weighted loss functions have been utilized to address this issue and improve
classification accuracy. Furthermore, hybrid models, combining XGBoost with Neural
Networks or SVM with Random Forests, have shown to enhance performance by integrating
strengths from different algorithms (Wang & Yao, 2013).
In addition, many studies have focused on interpretability, particularly for complex models
like Neural Networks and SVMs, which are difficult to explain. Rule-based models and rule
extraction techniques have been used to generate if-then rules, making these models more
transparent and helping in the decision-making process (Barakat & Bradley, 2010; Zięba et
al., 2014).
In summary, while traditional techniques like Logistic Regression and SVM have been
effective in loan prediction, modern methods such as Random Forest, XGBoost, and Neural
Networks are increasingly used, focusing on addressing class imbalance, improving model
accuracy, and enhancing interpretability.
Loan Prediction Methodology:
Collection of Data set

Feature selection using info gain of features

Train model on training data set

Test the model on testing data set

Result analysis