Credit Card Approval Prediction

Project Overview

This project aims to predict credit card approval using machine learning techniques. The goal is to build and evaluate various classification models to identify the best-performing model for accurate prediction.

Key Steps:

1. Data Preparation

   • Loading and merging credit card datasets (Credit_card.csv and Credit_card_label.csv).
   • Handling missing values by removing irrelevant columns and rows.
   • Checking for and removing duplicate entries.

2. Exploratory Data Analysis (EDA)

   • Calculating descriptive statistics to understand the data.
   • Analyzing the average income by gender and visualizing with a bar plot.
   • Visualizing the credit approval status distribution using a pie chart.
   • Exploring relationships between features with visualizations like box plots and violin plots.

3. Visualization

Average Income by Gender

Credit Approval Status

Annual Income By Marital Status

Birthday Count By Gender

Family Members By Marital Status

Annual Income By Education

4. Feature Engineering and Selection

   • Selecting relevant features for model building (Car_Owner, Propert_Owner, Annual_income, EDUCATION, label).
   • Converting categorical features to numerical representations using label encoding.

5. Model Building and Evaluation
   Splitting the dataset into training and testing sets.

   • Standardizing data using StandardScaler.
   • Training and evaluating various classification models:
     • Logistic Regression
     • K-Nearest Neighbors (KNN) with GridSearchCV for hyperparameter tuning
     • Support Vector Machine (SVM) with GridSearchCV
     • Decision Tree with GridSearchCV
     • Random Forest with GridSearchCV
     • AdaBoost with GridSearchCV
   • Evaluating model performance using accuracy score.

6. Results

Model	Accuracy Score
Logistic Regression	0.90
K-Nearest Neighbors (KNN)	0.60
Support Vector Classifier (SVC)	0.80
Decision Tree	0.80
Random Forest Classifier	0.85
AdaBoost	0.90

Best Performing Models:

The Logistic Regression and AdaBoost models achieved the highest accuracy of 0.90, making them the best-performing models for credit card approval prediction in this project.

---

Usage

1. Clone the repository:
2. Activate the virtual environment (if created):
3. Run the Jupyter Notebook:

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License.