A Model for Predicting Failures and Planning Maintenance of Bank ATMs Using Deep Learning

Abstract

The present study proposes a model for predicting failures and planning the maintenance of bank ATMs using deep learning. To achieve reliable results, the meta-synthesis method was selected as the primary approach for data analysis and modeling. This study is designed as an applied, descriptive, and qualitative research and was conducted using the meta-synthesis method. Initially, relevant articles on the research topic were selected, and after applying various screening processes, they were chosen for analysis. The research population includes all articles published in English and Persian journals from 2018 to 2024 that were accessible through reputable scientific databases. In this study, advanced machine learning and deep learning techniques were employed for failure prediction, infrastructure monitoring, and fraud detection, and the obtained results were compared and analyzed with previous research. This study aims to improve prediction processes and enhance the efficiency of various systems by examining the application of these techniques in different fields. The results indicate that utilizing these methods can significantly increase prediction accuracy and prevent unexpected failures, natural disasters, and financial fraud. Based on the research findings, it is recommended that organizations and industries adopt these technologies to optimize processes, increase efficiency, and improve decision-making. This study is particularly applicable in the fields of infrastructure management, financial security, and natural disaster prediction and can contribute to cost reduction and increased customer satisfaction.