Kaggle's Titanic with Pytorch and sci-kit learn
Data handling, cleaning and feature selection has been performed, before using machine learning techniques.
The notebook contained in this repository employs two different methods:
- A 3-layer neural network coded in Pytorch
- A second more "data scientist"'s method, to compare machine learning models from sklearn (and xgboost).
The second method has been written such that it can be used as a template for a lot of projects.
The final submission was made using the neural network's prediction, which shows the power of neural network.
The kaggle kernel can be found here https://www.kaggle.com/parthshxh/machine-learning-with-sklearn-and-pytorch