Leveraging Machine Learning to Forecast Trends in Cryptocurrency Exchange Markets

This research investigates the application of machine learning (ML) algorithms to enhance the predictability of movements in the cryptocurrency market, specifically examining changes in Bitcoin prices. Four ML models, Linear Regression (LR), Random Forest (RF), Gradient Boosting Machines (GBM), and Long Short-Term Memory (LSTM), were analyzed using historical data. The performance of these models was evaluated based on their accuracy, precision, and recall. The findings indicated that Random Forest surpassed the others with nearly perfect results in accuracy (0.99), precision (0.99), and recall (0.99). GBM demonstrated strong recall (0.99) but had lower accuracy (0.83) and precision (0.75), while Linear Regression showed commendable performance with accuracy (0.93) and precision (0.97). LSTM yielded the least favorable results. The study concludes that Random Forest is the most dependable model for predicting cryptocurrency prices, providing useful insights for traders and researchers navigating this highly volatile market.