App for predicting cryptocurrency price fluctuations with neural networks

• Main Author: Ho, Tristan Yue Ming
• Other Authors: Wong Liang Jie
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science
• Online Access: https://hdl.handle.net/10356/177341
• Institution: Nanyang Technological University

This report introduces a comprehensive analysis and implementation of the prediction of cryptocurrency using neural networks. The main focus will be bitcoin (BTC). The objective of this study was to develop a predictive model that is able to forecast future bitcoin prices based on historical price data. The predictive model employs Long Short Term Memory (LTSM) as well as Bidirectional LTSM networks. The code was made to fetch historical BTC price data from Binance API. Data was pre-processed then split into train and test sets to evaluate the performance of the models. LTSM and Bidirectional LTSM were trained on this data. The results showed both models outputting price predictions with the Bidirectional LTSM outperforming LTSM in terms of accuracy in prediction. Both models were plotted against actual price movements. In conclusion this study showed that Bidirectional LTSM has significant potential in predicting cryptocurrency prices. However, more research and enhanced engineering would be needed to improve the prediction accuracy as the nature of cryptocurrency is volatile.