AI-Driven Return Rate Prediction in Online Retail using Behavioral and Transactional Data

The explosive growth of e-commerce has created unprecedented problems in handling product returns, with online return rates frequently in excess of 30% of all sales. The present review is a comprehensive analysis of the state-of-the-art applications of artificial intelligence and machine learning methods in predicting return rates in online retail settings based on behavioral and transactional data. By structured review of recent literature and empirical testing on a dataset of 6,600 customer transactions, this paper integrates existing methodologies, determines important predictive features, and compares the performance of different machine learning methods. This paper’s findings illustrate that ensemble methods, specifically XGBoost and CatBoost, produce best results with F1-scores above 0.92, whereas behavioral features like session time, return history in the past, and transaction patterns prove to be important predictors. The review points out major shortcomings of existing research and suggests avenues for future research towards enhancing AI-powered return prediction functions in e-commerce.