Deep learning techniques business performance optimization in micro, small, and medium-sized enterprises: systematic review

Sampedro Guaman C.R.; Lengua M.A.C.; Rodriguez C.R.; Aguilar-Alonso I. · Bulletin of Electrical Engineering and Informatics (2026) · DOI: 10.11591/eei.v15i1.9819

Resumen

The application of deep learning is transforming how micro, small, and medium-sized enterprises (MSMEs) operate. By using data-driven insights, these firms overcome traditional analytical limitations and improve decision-making. This study explores factors influencing deep learning adoption in MSMEs, identifies effective strategies, and compares performance between companies that implement these methods and those that do not. The objective is to analyze the impact of deep learning on optimizing the performance of MSMEs. The methodology consisted of a scientific review following the preferred reporting items for systematic reviews and meta-analyses (PRISMA) system and a bibliometric analysis to map international contributions. The results show that techniques such as recurrent neural networks (RNNs), long short-term memory (LSTM) networks, transformers, and deep reinforcement learning (DRL) are crucial for marketing strategy prediction, customer experience personalization, and inventory management, leading to better return on investment (ROI), loyalty, and efficiency. Despite the potential benefits, there's still no enough research on how small businesses with limited resources use these methods and deal with issues like poor infrastructure and data access. Deep learning is essential for MSMEs' sustainability and competitiveness, even if there are challenges.

Instituciones: Universidad Nacional Mayor de San Marcos; Universidad Tecnológica del Perú; Universidad Nacional Tecnológica de Lima Sur