A Systematic Review of Causal Machine Learning in Business

Zevallos-Durand G.; Guevara-Ponce V.; Roque-Paredes O.; Iparraguirre-Villanueva O. · SSRG International Journal of Electronics and Communication Engineering (2025) · DOI: 10.14445/23488549/ijece-v12i11p103

Resumen

Causal inference focuses on understanding and explaining cause-and-effect relationships between variables, while Machine Learning (ML) primarily seeks to optimize prediction. Objective: To determine how machine learning algorithms are being used for causal inference in scientific research in the business field, identifying trends, gaps, and opportunities for methodological development. Method: A systematic review of the literature was conducted following PRISMA guidelines. Searches were performed in international scientific databases, initially identifying 1,998 articles. After applying inclusion and exclusion criteria related to thematic relevance, study type, full-text availability, and methodological quality, only 17 articles were selected for final analysis. Results: demonstrated an emerging use of the intersection between machine learning and causal inference in business research. The applications identified focus on estimating causal effects in marketing, analyzing consumer behavior, and studying the impact of business policies. However, limited adoption of algorithms specifically designed for causal inference was observed, with regularized regression models and matching methods being the most widely used techniques, while more advanced approaches, such as causal forests or doubly unbiased machine learning, were rare. The small number of articles (17) compared to the initial universe is an indicator of the existing gap. Conclusions: Finally, in the business world, the potential of integrating ML with causal inference has not yet been fully exploited.

Instituciones: Universidad Ricardo Palma; Universidad Tecnológica del Perú