Skip to content

Regression/mlp/ishwari

Niphade, Ishwari (PG/T - Comp Sci & Elec Eng) requested to merge regression/mlp/ishwari into main

This merge request adds the complete MLP regression pipeline for the Incident Management dataset. It includes data preprocessing, feature engineering, model training, hyperparameter tuning using RandomizedSearchCV, and evaluation.

Features Implemented

  • Feature selection based on Pearson correlation
  • Removal of collinear and low-importance features
  • MLP model training using Scikit-learn
  • Hyperparameter tuning with RandomizedSearchCV
  • Evaluation metrics: MAE, MSE, R²
  • Visualizations: distribution comparison, feature importance
  • Model saved in models/ folder for future reuse

Key Additions

  • notebooks/regression/mlp/ishwari_mlp.ipynb: Full experiment notebook
  • models/mlp_regressor_best.pkl: Saved best-performing model

Merge request reports