This repo contains a solution for the competition from Kaggle: House Prices - Advanced Regression Techniques.
The best model (XGBoost) achieved a score of 0.14. When trained with the data embedding from the autoencoder, the predictions are slightly improved.
According to the analysis of fedesoriano, a good and realistic model should be able to accomplish a score between 0.10 and 0.77, whilst a top model should score between 0.10 and 0.14.
- run.py --> End to end analysis.
- EDA.py --> Exploratory Data Analysis and data depuration.
- MLs.py --> Machine Learning implementations.
- Superlearner --> Stacked Machine Learning model implmementation.
- DLs.py --> Autoencoder Implementation.
- Numpy
- Pandas
- Matplotlib
- Seaborn
- Scikit-learn
- xgboost
- Tensorflow