Keywords: regression models, gradient boosting, variable reduction, feature engineering
We explore a variety of statistical techniques for predicting the sale price of homes using a base dataset from Kaggle. After consolidating the base dataset’s variables, we add additional features such as crime rates, public school ratings, and property taxes. Due to the range of years the dataset covers, normalization is used to counteract the effects of inflation in the given sale prices.Our method utilizes data mining techniques including advanced regression models, gradient boosting, variable reduction, and feature engineering. We explain the relationships among home features to predict final prices of homes.