This repository contains a data mining solution to a problem assigned to students of an Athens University of Economics & Bussiness. The project assignment unfortunatelly is restricted to students-only, but an outline can be found on the Kaggle Competition used for tracking progress of student teams.
The team members were the following:
The project was fullfilled using Random Forests for classification, which were programmatically optimized and tweaked to achieve sweetspot results on the classification. More meaningful features were also extracted from the dataset. Finally, using the Kaggle API, an auto-submitter was implemented.
Any code you find in this repository is under the GPL License. Please do not use our hard work without attribution. Also, if you are a class student, we advise you to implement your own solution to the problem. We also feel the need to highlight that this project was research-oriented, meaning we did not pick the best/most accurate Machine Learning algorithm, we picked the one we thought our knowledge would benefit the most from.
Special thanks to the Thinking Cup which provided us with a steady coffee flow throughout the project implementation.