salman104 / kaggle-house-prices-advanced-regression-techniques

Udacity capstone project: Kaggle competition on house prices prediction using advanced regression techniques

Kaggle House Prices: Advanced Regression Techniques

House Prices: Advanced Regression Techniques Competition on Kaggle

Shitao Wang submission version 1

Install

This project requires Python 2.7 and the following Python libraries installed:

• NumPy
• matplotlib
• seaborn
• scikit-learn
• XGBoost

You will also need to have software installed to run and execute an iPython Notebook

Code

All ipython notebook are used for data preprocessing, feature transforming and outlier detecting. All core scripts are in code folder, in which the ensemble learning script is in ensemble folder and base model script is in sing_model folder. All input data are in input folder and the detailed description of the data can be found in Kaggle.

Run

For a single model run, navigate to the /code/single_model/ and run the following commands: python base_model.py For a ensemble run, navigate to the /code/ensemble/ and run the following commands: python ensemble.py Make sure to change the data directory and the parameters accordingly before the model run.

Submission

Submission score on Kaggle leaderboard with different approaches.

FlowChart

Flow chart of the code.

Documentation

See ./doc/capstone_doc.pdf for detailed project documentation.