aastha985/Employee_Attrition_Prediction

Implements comparative evaluation of 9 supervised classification models (Logistic Regression, Random Forest, SVM, MLP, etc.) on the IBM HR attrition dataset across 6 preprocessing variants including class balancing techniques and PCA dimensionality reduction. Applies hyperparameter optimization via RandomSearchCV/GridSearchCV with 5-fold cross-validation and feature importance analysis to identify key attrition drivers. Achieves 99.2% accuracy using Random Forest with PCA and oversampling, emphasizing precision/recall/F1 metrics over accuracy due to class imbalance in the binary classification task.

No commits in the last 6 months.