Forecasting Online Shoppers Purchase Intentions with Cat Boost Classifier

Online shopping refers to the process of purchasing goods or services over the internet from e-commerce websites or online marketplaces allowing consumers to browse, select, and buy products without physically visiting a store.Users do not necessarily visit e-commerce sites with a buy purpose, but rather to browse and engage with things before leaving without generating cash for the firm. This is the challenge that every e-commerce faces. To solve this challenge it is necessary to predict the purchasing intentions of online shoppers for focusing on interested customers and generating income in the future.After knowing customer intention(buy or not buy) it is easy to generate individualized product recommendations, offer discounts, and perform numerous interventions to improve the overall customer experience.Predicting customer purchase behavior ahead of time is difficult. In this study,a proposed system is build to predict the online shoppers intention (buy or not buy) by analyzing different classification algorithms such as Random Forest, Decision Tree, Support Vector Machine, Gradient Boosting, Ada Boosting, LightGBM, Cat Boost. The experiments were performed on online shoppers intention dataset. Experimental results show that Cat Boost classifier with hyperparameter tuning can predict with highest accuracy, which is 98.73%.