MLP-NARX Bitcoin Price Prediction Model Integrating System Identification Modelling Principles

Bitcoin is a decentralized digital currency that enables people to exchange value without requiring third-party intermediary. Due its many advantages, it has received much interest from institutional and individual investors. Despite meteoric increase, the price of extremely volatile asset class as purely relies on supply demand. This presents an interesting opportunity create forecasting model. However, research papers in this area does not analyse residuals part resulting potentially biased models. In paper, we demonstrate System Identification (SI) residual analysis techniques our The Multi-Layer Perceptron (MLP) Nonlinear Autoregressive with Exogeneous Inputs (NARX) uses historical data several technical indicators predict future movements Bitcoin. Particle Swarm Optimization (PSO) algorithm was used find optimal parameters for model able one day ahead prediction test. successfully captured dynamics through tests performed residuals. It also proving randomness residuals, albeit some minor violations.