Exploring the Influence of Global Conflicts on Cryptocurrency Markets with RNN-Based Analysis

Cryptocurrency, a digital asset secured by cryptography, includes faster money transactions and innovative financial opportunities with its decentralized structure. Since it is a worldwide asset, it encounters challenges like fluctuations, geopolitical uncertainty, security vulnerabilities, and liquidity issues. So, it is difficult for traders to understand the crypto market and find lucrative possibilities that lead to navigating this volatile market to make it manageable. In our work, we chose Bitcoin, the first and most well-known cryptocurrency, to forecast the closing prices. We divided our whole dataset based on two geopolitical events (the Russia-Ukraine and the Israel-Palestine war) and applied the LSTM, GRU, BiLSTM, and BiGRU models from the RNNs to them. We evaluated each of our models on divisions of the dataset with measurement scales like MAE, MAPE, SMAPE, MSE, RMSE, and R2. The BiGRU models gave a higher performance in most of the divisions of datasets which leads BiGRU models as a stable model in geopolitical inconsistencies on Bitcoin’s closing prices. Our proposed system has the potential to serve traders and investors as a sophisticated Decision Support System (DSS) in the cryptocurrency industry.