• Asif Ali Ansari
• Sarita Gulia
• Rashmi Gupta

Cryptocurrency trading is challenging to forecast since the markets are mostly unpredictable. This work suggests a combined GARCH, LSTM, and XGBoost approach to estimate the price variations of BTC, ETH, DOGE, ADA, and WBTC. Technical indicators, lags, and a range of other distinctive Statistical features are applied to the framework to improve its capabilities in many temporal and market situations. The process involves regularly reviewing settings and also handles constantly updated data to adjust in real-time to new market changes. Experiments confirm that the new hybrid method always performs better than conventional models according to RMSE and MAE in predicting the volatility of multiple cryptocurrencies. The model’s strong results suggest it can be applied to algorithmic trading, making forecasts, and managing risks.

Cryptocurrency, Feature Engineering, GARCH, Hybrid Model, LSTM, Time Series Forecasting, Volatility Prediction.

"Hybrid Deep Learning Framework for Real-Time Cryptocurrency Volatility Prediction", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 5, page no.i103-i111, May-2025, Available :http://www.jetir.org/papers/JETIR2505911.pdf

"Hybrid Deep Learning Framework for Real-Time Cryptocurrency Volatility Prediction", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 5, page no. ppi103-i111, May-2025, Available at : http://www.jetir.org/papers/JETIR2505911.pdf