Data Cleaning​

• Exploratory Data Analysis & Visualization​
• Dealt with missing values​
• Log target variables for better distribution​
• Normalized data using skew and Box Cox transformations​
• Applied RobustScaler on number columns​
• Shifted outliers of prime attributes closer to the linear regression and used ZScore to drop outliers of other attributes.​

Hyperparameter tuning using Random Search​

Model building, stacking regression methods.​

• experimented with: xgBoost, Lasso, Ridge, SVR, LGBM, Gradient Boosting Regressor, Keras, Tensor Flow, ENet, KRR.​

Testing & validating the model​

Final prediction: -> Ridge, Lasso as meta regressors + SVR, LGBM and Gradient Boosting Regressor.​