Hybrid Deep Learning Architecture Combining Attention Mechanisms and Recurrent Neural Networks for Enhanced Time Series Analysis
Aditya Nagdiya, Vivek Kapoor, Vrinda Tokekar · Procedia Computer Science · 2025
In recent years, attention mechanisms have revolutionized various domains of deep learning, particularly in natural language processing and computer vision. However, their application in time series analysis remains relatively underexplored. This research proposes a novel hybrid deep learning architecture that integrates attention-based models with recurrent neural networks (RNNs) to improve the accuracy and interpretability of time series forecasting. The proposed architecture leverages the strengths of attention mechanisms to capture long-term dependencies and the sequential modeling capabilities of RNNs. Extensive experiments on benchmark datasets demonstrate the superiority of the proposed hybrid model over existing state-of-the-art methods.