Classifcation of data of an E-commerce website using ML Techniques

– Skill-development project under Machine Learning Practices course of IIT, Madras to strengthen knowledge of ML Techniques and Practices. – This was a project as well as a competition on Kaggle. Secured Rank 13 among 700+ participants. – Used Machine Learning for Classifcation of data collected from an e-commerce website with the objective to determine whether a user Purchases the product or not. – Achieved 75% F1 score using Stacking Classifer consisting of Support Vector Machine with RBF Kernal, Random Forest, Extra Trees, GradientBoost and Logistic Regression boosted by AdaBoost Algorithm through Quantile Data Transformation technique evaluated with 5-Fold Stratied Cross-Validation. – Also applied other algorithms including K-Nearest Neighbours(KNN) Classier, Decision Tree Classier, RidgeClassifer, NaiveBayes Classifers, Perceptron, Support Vector Machines Algorithm with Linear, Polynomial and RBF Kernal, XGBoost Classifer. – In addition to above algorithms also applied various Ensembling methods such as Bagging, Boosting (GradientBoost, Adaboost, XGBoost), Voting, Random Forest Classifcation using CART.