This project aimed to predict the approval status of loan applications using Support Vector Machines (SVM). By analyzing customer data and financial attributes, the model provides a reliable decision-making tool for financial institutions. 

• Methodology:

  • Data Preprocessing:

    • Cleaned and handled missing values in the dataset, focusing on features like income, loan amount, credit history, and dependents.

    • Applied one-hot encoding for categorical variables and standardized numerical features for optimal model performance.

  • Feature Engineering:

    • Selected significant features using statistical methods and correlation analysis to enhance prediction accuracy.

  • Model Development:

    • Implemented an SVM classifier with a linear kernel to classify loan applications as Approved or Rejected.

    • Evaluated performance using metrics such as accuracy, precision, recall, and F1-score.

  • Hyperparameter Tuning:

    • Used grid search to optimize hyperparameters like the regularization parameter (C) and kernel choice for improved model generalization.

• Key Results and Insights:

  • Achieved an accuracy of XX% and a balanced precision-recall score, ensuring minimal bias in predictions.

  • Identified key predictors such as credit history and applicant income as critical factors influencing loan approval.

  • Demonstrated the effectiveness of SVM in handling binary classification tasks with high-dimensional data.

• Skills and Tools:

  • Machine Learning (SVM), Python (Pandas, Scikit-learn, Matplotlib, Seaborn), Data Cleaning, Feature Engineering, and Model Evaluation.

This project highlights the use of machine learning in the financial domain, showcasing the ability to develop robust models that aid in automating loan approval processes while minimizing risk.

Github link: https://github.com/sm98code/Loan-status-prediction-using-svm.git