Predict house prices in Ilinois
- Description.ipynb has the logging and the plots used for analysis and processing of the data
- preprocessing.ipynb has the working version of the data preprocessing
- model.ipynb contains the final model (the last one in the notebook) along with a some the others that were tested
Run the preprocessing notebook, which takes the train.csv file and processes it. This script creates multiple files, each representing different preprocessing stages. The final output is a version containing the most representative and treated features. It also handles the preprocessing for the test.csv since it has to follow the same rules.
These are the algorithms we tried with grid search: Linear regression Random forest Extra trees XGB boost
Regretably, the most important plot of our data did not render properly in Description.ipynb, so we are appending it here. The following is the plot of feature importance:
