Amazon Sales Prediction Model Using ML Algorithms

Abstract

The e-commerce landscape has rapidly evolved, becoming a cornerstone of modern business with its digital paradigm for buying and selling goods or services. This dynamic and expansive platform, facilitated through websites and applications, transcends geographical boundaries, offering unprecedented access to a global consumer base. In this context, the ability to predict product sales has emerged as a critical facet for businesses operating in the e-commerce domain. The digital marketplace is characterized by fierce competition and rapid changes in consumer preferences. As businesses strive to optimize their operations, enhance profitability, and stay ahead of the curve, the predictive modelling of sales becomes invaluable. By forecasting sales, businesses can strategically adjust pricing strategies, manage inventory effectively, and tailor marketing efforts based on anticipated demand. The three fundamental factors—Price, Review, and Rating—constitute the core elements influencing consumer decisions in e-commerce. Understanding the intricate interplay of these variables is essential for businesses seeking a competitive edge. This research employs the Design of Experiment (DOE) technique to meticulously examine the relationship between Price, Review, Rating, and product sales. Leveraging a diverse dataset from Amazon, the study provides a comprehensive analysis of how these factors synergistically impact consumer behaviour and purchasing patterns. The resulting mathematical model demonstrates a promising capacity to forecast product sales accurately. In essence, the significance of predicting sales in e-commerce lies in its transformative potential for businesses, offering actionable insights that pave the way for informed decision-making, enhanced customer satisfaction, and sustained competitiveness in the ever-evolving digital marketplace. Keywords: E-commerce, Sales, Marketplace, DOE, Mathematical Model, Forecast.