An Efficient Predictive Analysis Model of Customer Purchase Behavior using Random Forest and XGBoost Algorithm

Abstract

Predictive Analytics is a bough of the advanced analytics which has been used to make predictions about unknown future events. It uses many techniques from Data Mining and Statistics Modelling to analyze the current data. In Statistics Modeling, Regression Analysis algorithms are some of the most popular processes used in Machine Learning Models. Random Forest is a supervised learning algorithm which uses Ensemble Learning method to take advantage of Bootstrap Aggregating. XGBoost is a scalable & accurate implementation of Gradient Boosting Machines (GBMs). It has been proved to push the limits of computing power. It is built & developed for the sole purpose of model performance and computational speed. Customers are the basis for growth of any type of business. In the study of sales and purchase, it is vital & crucial to be able to predict the amount of purchase or sales to increase benefit by catering from specific products to specific demographics. Our prediction analysis model can effectively help to improve the performance and increase the profit margin. Moreover, it can generalize the prediction of purchase or sales figures in any market which depends on the customers' past purchase pattern or behavior.