A Stacking Ensemble of Multi Layer Perceptrons to Predict Online Shoppers' Purchasing Intention

Abstract

With the rapid development of the internet, the field of E-Commerce has seen tremendous growth. The easy accessibility of viewing and purchasing products, and having it delivered to your doorstep is what makes E-Commerce extremely successful. E-Commerce also caters to almost every possible field, ranging from electronics to fashion to groceries. The number of visitors on E-Commerce websites is growing at a rapid pace, but the number of purchases remains constant. A novel stacking ensemble system has been proposed, which makes use of Multi-Layer Perceptron's to detect the intention of a user on whether a product would be purchased or not. ‘Online Shoppers Purchasing Intention’ Dataset has been used. The proposed stacking ensemble model achieved an accuracy of 94% in predicting whether a user will purchase a product or not, based on the session details. To evaluate the proposed system, it is compared to over 15 classification algorithms, as well as existing systems that make use of the dataset. The results obtained show that the proposed stacking classifier outperforms the various classification algorithms as well as existing systems that make use of the dataset.