Cryptocurrency Price Forecasting Using Deep Learning Model with Technical Indicators

Abstract

This research explores the application of an Attention-based GRU model for predicting Bitcoin price movements. Historical data from Yahoo Finance, along with technical indicators such as the Stochastic Oscillator (KD index), Relative Strength Index (RSI), and Moving Average Convergence Divergence (MACD), were used to enhance prediction accuracy. The model was designed to focus on key time-series patterns, with the attention mechanism improving its ability to capture crucial market signals. Results show that incorporating these indicators improved performance, with the MACD-enhanced GRU model achieving an accuracy of 75%. The model's effectiveness in volatile cryptocurrency markets highlights the advantages of deep learning models, combined with technical indicators, for accurate financial forecasting.