We are a group of two, including Yi Yang, Di Niu. We will contribute to this repository on Kaggle House Price Competition and Digit Recongnizer Competition.
This github repository shows the process of how we are trying to improve the model performance for both Kaggle housing price and digital recognizer contests from time to time. We have used Professor Patrick Hall's H2O ipython notebook as references to start our trial. We have also looked up Kaggle's discussion board and coding websites to improve our understanding of deep learning.
The final results show a great improvement in terms of Kaggle leaderboard ranking. We are able to make top 10% in both Kaggle contests, although they are both intro-level contests and no actual ranking points are awarded. We appreciate the learning experience and gains in this short 6 weeks course. As for the detail ipython notebook, please refer to those two files: Housingprice_Final.ipynb and Digital_Recognizer_CNN_Final.ipynb.
Thank you, Professor Patrick and Lisa!