• Priya Sidhu
• Dr. Himanshu Aggarwal
• Dr. Madan Lal

The key challenges that still persist is the inability to forecast accurately the prices of stocks because of the volatility of financial markets as well as non-linearity of the phenomenon. This paper suggests a hybrid deep learning model that will combine Discrete Wavelet Transform (DWT) to assuage the signal; fuzzy logic-controlled feature extraction to model the uncertainties; and a Bidirectional Gated Recurrent Unit (BiGRU) framework with Attention mechanism to learn the temporal domain efficiently. In order to maximise the performance of a model even further, Optuna was utilized to auto-tune the hyper-parameters and ensemble averaging was applied to stop fluctuations in predictions. The proposed DWT-Fuzzy-BiGRU-Attention-Optuna model was applied to a pair of Indian market equities Asian Paints and Maruti Suzuki obtained in Yahoo Finance. The results suggest that the suggested model is reasonable in terms of capturing the complexities of time-related dependencies, as well as providing a high degree of predictions accuracy in various market conditions.

"A Multi-Resolution Fuzzy–BiGRU–Attention Ensemble Tuned with Optuna for Robust Stock Forecasting", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 10, page no.e700-e706, October-2025, Available :http://www.jetir.org/papers/JETIR2510483.pdf

"A Multi-Resolution Fuzzy–BiGRU–Attention Ensemble Tuned with Optuna for Robust Stock Forecasting", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 10, page no. ppe700-e706, October-2025, Available at : http://www.jetir.org/papers/JETIR2510483.pdf