Currently in progress.
This project is a survey of several common forecasting algorithms, including various time-series and sequential neural network models, as applied to the cryptocurrency market. The goal is to explore the mathematical and programming foundations of these algorithms for more advanced market modeling.
The script Time Series Analysis builds and compares the following models with Bitcoin price data: AR, MA, ARMA, ARIMA, ARIMAX, SARIMA, SARIMAX, and a Martingale Model. For the exogenous variables, we used the technical indicator Relative Strength Index.