The aim of this project is to predict house prices using one basic machine learning algorithm, Linear Regression, and one advanced algorithm, Random Forest. We will also use regression with regularization such as Ridge and Lasso to try to improve our prediction accuracy.
The Kaggle House Prices datasets can be downloaded here.
Random Forest was found to be the better model for predicting house prices. It out performed the regession algorithms with performance accuracy of 85% using R-squared metric. The most important predictor was the overall quality of a house, following the size of above ground living area and the total basement square footage.
This project is a first pass to get us quickly to a reasonable good model prototype.