Algorithmic trading with cryptocurrencies - what is the optimal modelling design for bitcoin price and trend prediction

Since its inception in 2009, Bitcoin has gained popularity and importance in financial markets. The Bitcoin price is highly volatile entailing high risk and chances of high returns for traders. This work is part of a work project that defines a holistic approach to build an intraday Bitcoin trading...

ver descrição completa

Detalhes bibliográficos
Autor principal: Nolden, Jannik (author)
Formato: masterThesis
Idioma:eng
Publicado em: 2022
Assuntos:
Forecasting
Machine learning
Business analytics
Cryptocurrency
Bitcoin
Deep learning
Price prediction
Algorithmic trading
Time-series forecasting
Domínio/Área Científica::Ciências Sociais::Economia e Gestão
Texto completo:http://hdl.handle.net/10362/142539
País:Portugal
Oai:oai:run.unl.pt:10362/142539
Descrição
Resumo:Since its inception in 2009, Bitcoin has gained popularity and importance in financial markets. The Bitcoin price is highly volatile entailing high risk and chances of high returns for traders. This work is part of a work project that defines a holistic approach to build an intraday Bitcoin trading algorithm based on predictive analysis of ML models. In this work, the results show that LSTM yields the best prediction performance for Bitcoin price prediction. GRU, LSTM and RNN demonstrate the best performance for Bitcoin trend prediction in 1h, 2h and 3h,respectively.

Registros relacionados