Abstract: The growth of e-commerce has revolutionized the way consumers purchase Fast-Moving Consumer Goods (FMCG). Accurate sales prediction in e-commerce platforms is vital for inventory management, supply chain optimization, and business success. This abstract presents a sales prediction model designed for FMCG products in e-commerce. It employs Machine Learning (ML) techniques to analyze historical sales data, product attributes, customer behavior, and external factors for predicting sales trends. Furthermore, this research recognizes the significance of competitor analysis in the sales prediction framework. By integrating competitor data into the model, businesses can assess their market share, pricing strategies, promotions, and customer preferences. The model aims to provide a comprehensive understanding of competitors, enabling businesses to make informed decisions and gain a competitive advantage. This study focuses on predicting sales using an ML algorithm trained on Best Seller Data (BSR) data. The aim is to develop a model that accurately predicts sales by leveraging the relationships captured in the BSR data. Implemented Multiple Linear Regression along with three regularization techniques: Ridge Regression, Lasso Regression, and Elastic Net Regression. Among the regularization methods, Elastic Net Regression yielded the highest accuracy of 90.10%. This indicates that the Elastic Net Regression model was able to capture the relationships between the predictor variables and the target variable more effectively compared to the other regression models.

Presented in: International Conference on Computing and Machine Learning (CML-2024)

Published in: Lecture Notes in Networks and Systems, vol 1144. Springer