Predictive Effectiveness of Machine Learning and Traditional Models in Production and Sales: A Systematic Literature Review

Farroñan Soplapuco C.N.; Yalico Fernández M.A. · Proceedings of the LACCEI international Multi-conference for Engineering, Education and Technology (2025) · DOI: 10.18687/leird2025.1.1.422

Resumen

In recent years, the application of Machine Learning (ML) and Deep Learning (DL) techniques in sales forecasting has gained significant relevance as a strategic tool to optimize business processes and decision-making. This Systematic Literature Review (SLR) aims to identify the most widely used models and assess their effectiveness in sales estimation across various commercial settings. Following the PRISMA methodology, fifty-two academic articles published between 2022 and 2025 were analyzed. The results indicate that the most commonly employed models are Random Forest, XGBoost, LSTM, and CNN, all of which outperform traditional methods such as ARIMA and linear regression. It is noteworthy that DL techniques and hybrid models achieve R2 values above 90% and mean absolute percentage errors (MAPE) below 10%, confirming their effectiveness in multivariable and dynamic contexts. © LEIRD 2025.All rights reserved.

Instituciones: Universidad Tecnológica del Perú