kaggle-rossmann competition
Prediction model for Kaggle/Rossmann competition. Contact me on my linkedin if you need more information.
You can find out more about this competition at: https://www.kaggle.com/c/rossmann-store-sales
steps
- clone repo
- install python, pandas, numpy, scipy/scikit, keras, hdf5, ...
- run the ./load_data_hdf5.py in src/data directory
- run the model from models directory for training
- If you want to experiment with lstm then you need to edit and run create_data.py. You should edit this file and file lstm_simple.py and specify how long sequences should you want to use. This model is not complete, i don't decide yet how to make predictions for test. I have also currently problem with new version of Keras library for LSTM.
folders
/data : datasets
/src/data : handling dataset, missing values mostly with pandas
/src/models : models for prediction
/src/models/results : results