A research team from the Oman College of Management and Technology (OCMT) has successfully developed a Feedback-Matching Neural Network (FMNN) to enhance the accuracy of time series forecasting. This neural network integrates the Feedback-Matching Algorithm (FMA) within a deep learning framework, enabling the recognition of recurring patterns in data and improving prediction precision.
The research team includes Dr. Hazem Migdady, Mr. Louay Al Nuaimy, and Dr. Mastan Muhammad from the Department of Computer Science and Management Information Systems. Their work was presented at the 10th International Conference on Information and Communication Technology (ICICT 2025), held on February 20, 2025. The conference, sponsored by Springer and organized by the Global Knowledge Research Foundation, is a prestigious global platform for discussing advancements in Information and Communication Technology.
The FMNN model marks a significant advancement in time series forecasting, demonstrating superior performance in fields such as finance, weather prediction, and energy management. It has outperformed traditional statistical models and modern neural networks in predictive accuracy. This initiative reflects OCMT’s commitment to advancing artificial intelligence and data analysis, fostering scientific innovation, and providing impactful real-world applications.
Development of a Neural Network for Feedback-Matching in Time Series Forecasting