E-commerce platforms are a primary place for people to find, compare, and ultimately purchase products. They employ Machine Learning (ML), Business Intelligence (BI), mathematical formalism, and artificial intelligence (AI) to generate valuable knowledge about customer behavior, bringing benefits for both customers themselves and sellers. The state-of-the-art in this area does not include a comprehensive and up-to-date survey that explores the most common goals of e-commerce-related studies and the suitable ML techniques and frameworks for particular cases. In this context, we introduce a systematic literature review that revisits recent initiatives to employ ML techniques on different e-commerce scenarios. The contributions to the state-of-the-art are twofold: (i) a comprehensive review of ML methods and their relationship with the target goals of e-commerce platforms, including impact on profit growth; (ii) a novel taxonomy to reorganize ML-based e-commerce initiatives, which helps researchers to compare and classify efforts in this evolving area. This comprehensive literature review enables researchers and e-commerce administrators to conduct innovation projects better and redirect budget and human resource efforts.

电子商务平台是人们寻找、比较并最终购买产品的主要场所。他们采用机器学习 (ML)、商业智能 (BI)、数学形式主义和人工智能 (AI) 来生成有关客户行为的宝贵知识，为客户和卖家带来好处。该领域的最新技术不包括探索电子商务相关研究的最常见目标以及适用于特定案例的 ML 技术和框架的全面和最新调查。在这种情况下，我们引入了系统的文献综述，回顾了最近在不同电子商务场景中采用 ML 技术的举措。对最先进技术的贡献有两个方面：(i) 对 ML 方法及其与电子商务平台目标目标的关系的全面审查，包括对利润增长的影响；(ii) 一种新的分类法，用于重组基于 ML 的电子商务计划，这有助于研究人员在这个不断发展的领域中比较和分类工作。这种全面的文献综述使研究人员和电子商务管理员能够更好地开展创新项目并重新调整预算和人力资源工作。