Machine Learning and Bitcoin price prediction
Huge amounts of available data and rapid growth in computation power have changed the world where we live and conduct our business in a way that was just a few years ago unimaginable. The financial industry is trying to adapt to a new, fast-changing financial world by using innovative technologies to create new or upgrade existing business models to create financial innovative companies – FinTech. In this work, we are focused on these two sides of FinTech, where one side is represented by machine learning as innovative technology and the other side by Bitcoin as a representative of a new innovative business model – cryptocurrency. The main aim of this work is to find out if the past performance could be used as an indicator of future behavior. The proposed approach is implemented in Python by creating a single-point LSTM model that represents a machine learning algorithm that will be predicting bitcoin prices based only on historical daily closing prices.
JEL code: C22, C53, G15,
Keywords: machine learning, LSTM, time series, bitcoin, FinTech