Recent Machine-Learning-Driven Developments in E-Commerce: Current Challenges and Future Perspectives

Abstract

E-commerce platforms are becoming a primary destination for individuals to find, compare, and eventually buy goods. They use machine learning (ML), business intelligence, and mathematical formalization to derive meaningful knowledge about consumer behavior, purchase prediction, personalization, inventory management, fraud detection, and other advantages for both customers and businesses. This paper will analyze a state-of-the-art development in e-commerce, leveraging ML algorithms or techniques that suit a particular area of business improvement. In this context, a literature review will be performed to revisit recent developments in utilizing ML in various e-commerce areas. The contribution to the state-of-the-art are threefold: (i) a comprehensive review of ML utilization in different e-commerce scenarios; (ii) the current challenges of ML implementation in the e-commerce sector and areas that still require improvement; and (iii) future directions regarding ML development related to e-commerce.