Hybrid Machine Learning Method for Product Sales Forecasting in E-Commerce

The term “e-commerce” refers to the practice of buying and selling products using the internet. E-commerce was created so that consumers wouldn't have to leave the comfort of their homes to purchase something. So, they can buy it online and have it delivered to their house in a few days. Online retailing has recently developed in a favorable environment because of the robust backing of national legislation. This year's pandemic has made the e-commerce sector's contribution to the growth of the country's economy more noticeable. The quantity and effectiveness of e-commerce networks and e-commerce businesses are growing under such circumstances. A network must've been able to better match user demands and perform admirably in all areas of collaboration and administration if it hopes to keep its competitive edge. Accurately predicting the sales volume of e-commerce networks at this time is crucial. This article presents a novel hybrid machine learning (ML) algorithm to accurately predict revenue growth for an online store. A hybrid cat-XGBoost (HC-XG) algorithm is suggested to predict the sale. The main goal of the proposed work is to forecast e-commerce sales, thus i the final selling value is included in the solution since it has a significant impact on sales in extremely competitive and price-sensitive contexts like e-commerce. We next continue with the development of the suggested solution and evaluate it for accuracy, error, and overall efficiency, and compare it to some competitor solutions. Thus, the proposed approach is compared with other existing models for sales forecasting and demonstrate that it provides the best overall performance in product sales forecasting.