A Predictive Analytics Model for Maximising Profit in e-commerce Companies

Abstract

Abstract

Predictive analytics have become a crucial aspect for many e-commerce companies. The primary objective of this work is to apply and compare different machine learning techniques to build an efficient and accurate predictive model for the analysis and identification of customers who are expected to purchase again in the next 90 days. The main performance goal of the predictive model is to maximize profit for the e-commerce company by providing coupons to non-returning customers so that they are tempted to make follow up purchases.

Keywords: SMOTE, xgboost, caret, ROC, AUC, cost matrix

Cite this Article

Sarjo Das, Parth Singh,Gaurav Puri. A Predictive Analytics Model for Maximising Profit in e-commerce Companies. E - Commerce for Future & Trends. 2017; 4(2): 19–32p.