Quantile Regression: House Price Split

June 20, 2015

Report Abuse
Linear regression models predict the mean of the target column given input features. In some applications such as price prediction we are interested in predicting the range or entire distribution of the target column instead of a single estimate. In this experiment, we attempt to predict the 25th, 50th and 75th percentiles for the price for an automobile given is other attributes. We use a tree-based quantile estimator that can handle non-gaussian distributions.