SVM (Support Vector Machine) - Banking Credit Card Eligibility Prediction

Overview

This project utilizes a Support Vector Machine (SVM) classification model to predict customer eligibility for banking credit cards. It employs various data preprocessing techniques, visualizations, and model optimization strategies to enhance prediction accuracy.

Project Goals

• Accurately classify customers based on their eligibility for credit cards.
• Explore and visualize data relationships to understand customer attributes influencing eligibility.
• Optimize the SVM model using hyperparameter tuning and handling data imbalance effectively.

Steps Included

1. Exploratory Data Analysis (EDA)

• Data exploration and cleaning
• Identifying key features and their impact

2. Data Visualization

• Correlation analysis using Heatmap
• Understanding feature relationships and data distribution

3. Data Scaling

• Standardization to ensure optimal model performance

4. Model Implementation

• SVM (Support Vector Machine) classification

5. Hyperparameter Optimization

• Searching optimal hyperparameters through sampling techniques to improve model accuracy

6. Handling Class Imbalance

• Implementation of SMOTE (Synthetic Minority Oversampling Technique)
• Addressing data imbalance to enhance predictive performance

7. Post-SMOTE Hyperparameter Optimization

• Further refining the SVM model post-balancing for improved accuracy

How to Run

To run this project locally:

1. Clone the repository:

   git clone https://github.com/azizzoaib786/banking-dataset-prediction-svm.git

2. Install the required dependencies:

   pip install -r requirements.txt

3. Open and run the notebook:

   • Open banking-dataset-prediction-svm.ipynb in Jupyter Notebook or JupyterLab.
   • Execute all cells (Run All).

Requirements

• Python (3.7 or higher recommended)
• scikit-learn
• pandas
• numpy
• matplotlib
• seaborn
• imbalanced-learn
• Jupyter Notebook/JupyterLab

Contributions

Contributions are encouraged! Feel free to fork the repository, make improvements, and submit a pull request.

Contact

• Email: [email protected]

License

This project is licensed under the MIT License.