فهرست نشریات علمی دسترسی آزاد ایران

Predicting cryptocurrency price trends is crucial for helping investors make better decisions, avoid losses, manage risks, and increase profits. Unlike traditional financial markets, cryptocurrency markets are known for their high and unpredictable volatility. This intense price fluctuation can greatly impact the accuracy of prediction models, making it necessary to consider volatility first.To address this challenge, it is essential to model price volatility before attempting to predict prices. In this study, a hybrid GARCH-LSTM method was used to improve prediction accuracy. GARCH is effective in analyzing price volatility, while LSTM excels at processing time-series data and capturing complex patterns. The analysis was conducted over a two-year period, from August 2, 2021, to August 2, 2023, showing how this combined approach can tackle the unique challenges of forecasting in the volatile cryptocurrency market. According to the results, the mean squared prediction error for Bitcoin was 0.0006, with the maximum prediction error being 0.002. In other words, the predictive power of this hybrid model is 99.0%, indicating a high reliability of the estimated results with this method.