Introduction

This project was for the course Machine Learning Practice.

The project was in fact a private Kaggle competition allowed only for the degree students who had taken Machine Learning Practice that term.

Problem statement - Given the historical data of visitors to an ecommerce site, classify test visitors as whether they will purchase something from the site or not.

Since this course taught us how to implement all machine learning techniques including preprocessing, feature engineering and modelling through scikit-learn framework, this project was meant to test our practical skills in that regard.

I achieved a classification accuracy of 0.82 and obtained second place in the leaderboard.

This is the notebook I finally submitted in the competition; of course I had made several other notebooks for intermediate steps and experiments: Notebook