Credit Card Issuance Prediction Model

Overview

This project implements a Supervised Machine Learning Model to predict the optimal conditions under which a bank can maximize profit by issuing credit cards to individuals. The project uses various predictive models, including Random Forest, Decision Tree, and Linear Regression, to determine whether issuing a credit card to a specific customer is a profitable decision or not.

The dataset for this project is sourced from the UCI Machine Learning Repository and includes data related to credit card fraud. The model predicts whether the bank should issue a credit card to a customer by analyzing historical data.

Features

• Supervised Learning: Utilizes labeled data to train the model.
• Predictive Models: Implements Random Forest, Decision Tree, and Linear Regression.
• Visualization: Includes decision tree visualization to provide insights into the model's decision-making process.
• Fraud Detection: Incorporates data analysis to assess credit card fraud likelihood.

Dataset

The dataset used in this project is sourced from the UCI Machine Learning Repository. It includes customer information, transaction history, and fraud indicators. The dataset is preprocessed to handle missing values and normalize features for optimal model performance.

Models Used

1. Random Forest: Provides robust and accurate predictions by creating an ensemble of decision trees.
2. Decision Tree: Trained to visualize decision-making and understand key factors influencing predictions.
3. Linear Regression: Explores relationships between features and the target variable for baseline predictions.

Project Workflow

1. Data Preprocessing:

   • Cleaned the dataset.
   • Handled missing values and performed feature scaling.
   • Split data into training and testing sets.

2. Model Training:

   • Trained models (Decision Tree, Random Forest, and Linear Regression) using the training dataset.
   • Evaluated performance using metrics like accuracy, precision, recall, and F1-score.

3. Visualization:

   • Visualized the Decision Tree to interpret the model's predictions.

4. Prediction:

   • Predicted whether the bank should issue a credit card to a specific customer.
   • Assessed the risk of fraud and potential profit for each case.

Results

• Model Accuracy: Achieved high accuracy in predicting profitable credit card issuance decisions.
• Decision Tree Visualization: Offers insights into the factors influencing model decisions.