Project Info: Rahuri Finance requires building an AI model to predict the "Loan Payment Failure" tendency (a binary classification problem) for a given loan application. Additionally, we need to identify any bias towards specific attributes.
Project Steps:
(a) Preprocessing:
Detail the preprocessing steps, including handling missing values, plotting graphs, data discretization, normalization, and data encoding.
(b) Initial Data Exploration:
Utilize techniques like association rule mining and clustering for initial data exploration to identify potential biases. Report any findings.
(c) Building the Classification Model:
Explore various classification techniques (e.g., Decision Trees, Neural Networks, Naïve Bayes, SVM, and Ensemble methods). Evaluate training and validation errors to determine model suitability. Use cross-validation for model assessment and perform hyperparameter tuning with tools like GridSearchCV. Report the final cross-validated accuracies and optimal hyperparameters for each classification model. Tools and Libraries:
Python Libraries: Pandas, NumPy, Matplotlib, and Scikit-learn.