Original Research Open access
Advanced Real-Time Anomaly Detection and Predictive Trend Modelling in Smart Systems using Deep Belief Networks Architectures
PatternIQ Mining 2025;2(1):97–107 · https://doi.org/10.70023/sahd/250209
Abstract
Intelligent systems are becoming more data-intensive and complicated, so it's essential to have reliable methods to recognize anomalies immediately and monitor trends to keep the system running well. The present research delves into using the DeepSense Framework, built up on top of Deep Belief Networks (DBNs), to create a system that can change with changing settings and recognize and anticipate anomalies accurately. DBNs are generally used to examine data collections with several dimensions because of their multilayered, unsupervised prior instruction and fine-tuning capacities. The framework is built to capture complicated, unpredictable relationships amongst the collected information to discover minor abnormalities and anticipate patterns in the future. The DeepSense Framework successfully reduces the amount of false positives while keeping the identification of anomalies responsiveness substantial, according to an experimental assessment. The structure also outperforms more traditional quantitative and neural network methods when it comes to trend predictions.
Keywords :
Anomaly detection, predictive modelling, DeepSense Framework, Deep Belief Networks, intelligent systems.
How to cite
alkato, A. A., & sakhnin, Y. (2025). Advanced Real-Time Anomaly Detection and Predictive Trend Modelling in Smart Systems using Deep Belief Networks Architectures. PatternIQ Mining, 2(1), 97–107. https://doi.org/10.70023/sahd/250209
Copyright and licence
© 2025 The Author(s). Published by Sahara Digital Publications. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) licence. This licence allows others to copy, redistribute, remix, adapt and build upon the work for non-commercial purposes, provided that the authors and the original publication are credited. The authors retain copyright. See Copyright and Licensing.