House price prediction for Delhi using XGBoost

This project aimed to predict the volatile house prices using various machine learning algorithms with special attention to XGBoost regression. The dataset went through aggressive preprocessing in the form of missing value handling, categorical features and encoding, removal of duplicates and outliers, and log transformation to normalize the target variable. Several different models like Linear Regression, Random Forest Regressor and XG Boost Regressor were used to check for differences in performances. Out of all the models reviewed XGBoost out performed everyone else in terms of accuracy, obviously optimise by hyper tuning parameters. Model performance was measured in terms of RMSE, MAE, R2 and adjusted R2 value. The research raises the importance of data transformation and comparative model analysis to achieve greater predictive accuracy for the pricing of real estate.