spaceship-titanic

This repo contains my submissions to the Spaceship Titanic competition on Kaggle.

ibm-deep-learning-final-project

This project employs different types of Artificial Neural Networks to achieve a decent result in the competition. It uses Python, Jupyter Notebook, keras, scikit-learn, pandas, numpy, matplotlib, seaborn. This project contains:

• initial data processing.
• constructing 4 types of neural networks using different depths, Dropout and Batch Normalization.
• hyperparameter search for the neural networks with K-fold cross validation.
• predicting using standalone models and ensembles.

The best results were achieved using ensembles of neural networks with Batch Normalization (see below).

The full code and detailed summary for the project can be found here.