🏦 Housing Finance Loan Approval Prediction

This project predicts whether a housing loan application will be approved or rejected using Logistic Regression. It leverages financial and demographic data to help financial institutions make data-driven decisions.

• Analyze housing finance dataset
• Perform data preprocessing & feature engineering
• Build Logistic Regression model
• Evaluate model using accuracy & ROC curve

The dataset includes key attributes such as:

• 💰 Income to Asset Ratio (IAR)
• 📉 Loan to Value Ratio (LTV)
• 📊 Fixed Obligation to Income Ratio (FOIR)
• 🏢 Employer Type
• 🏦 Bank Savings
• 💍 Marital Status
• 🎯 Loan Decision (Target Variable)

• Language: R

• IDE: RStudio

• Libraries:

  • fastDummies
  • corrplot
  • caTools
  • car
  • ROCR

https://1drv.ms/x/c/f483042b9735aab9/IQCcLU9hPxEoSKeUcyso96AtAXgJrAD9s0hlR95DtLMv7L4?e=NLESkm

• Converted categorical variables into numerical format
• Used label encoding & dummy variables

• Correlation analysis to detect relationships & multicollinearity

• 70% Training Data
• 30% Testing Data

• Built Logistic Regression using glm() function

• Confusion Matrix
• ROC Curve
• Accuracy Score

• ✅ Achieved ~74% accuracy

• 📊 Key influencing features:

  • Income to Asset Ratio
  • Loan to Value Ratio
  • FOIR
  • Employer Type
  • Bank Savings

• Correlation Matrix

• 

• Confusion Matrix

• 

• ROC Curve

• 

# Install required libraries
install.packages(c("fastDummies","corrplot","caTools","car","ROCR"))

# Load dataset
data <- read.csv("your_dataset.csv")

# Run model (example)
model <- glm(Decision ~ ., data=data, family=binomial)
summary(model)

• Logistic Regression effectively predicts loan approval decisions
• Financial ratios play a major role in decision-making
• Machine learning improves accuracy and reduces manual effort

• Use advanced models (Random Forest, XGBoost)
• Hyperparameter tuning
• Deploy as a web app (Shiny)
• Improve accuracy with feature selection

Dhathri Narne 📊 Data Analyst | Machine Learning Enthusiast

Give it a ⭐ on GitHub!