| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.36 MB | Adobe PDF |
Advisor(s)
Abstract(s)
O mercado das criptomoedas é atualmente uma das áreas de maior interesse de investimento,
atraindo investidores desde os mais experientes até aos mais casuais, e embora possa
proporcionar elevadas rentabilidades também representam um elevado risco devido à sua
elevada volatilidade.
Neste contexto a inteligência artificial, sobretudo através dos algoritmos de aprendizagem
profunda e aprendizagem máquina tem assumido um papel preponderante no
desenvolvimento de aplicações que permitam aconselhar investidores, tentando maximizar os
retornos e reduzir riscos de investimento.
O presente trabalho propõe um sistema de previsão do preço de fecho de dez das principais
criptomoedas atualmente presentes no mercado, disponível numa aplicação web, capaz de
efetuar previsões de uma até quatro horas. Para tal foram analisados e testados diferentes
modelos com diferentes algoritmos de aprendizagem máquina e aprendizagem profunda, como
as Redes Neuronais Recorrentes, algoritmos de análise temporal como ARIMA e até alguns
algoritmos de regressão mais convencionais.
Para comparação dos algoritmos, foram usados os registos ao minuto dos preços da Bitcoin
relativos ao período de 30 dias, para a previsão a 60 minutos, e o modelo que apresentou
melhor desempenho foi o de Redes Neuronais GRU, usando todos os atributos das cotações,
obtendo um MAPE = 0,09%, MSE=5954,89, RMSE=77,17 e MAE=60,20 .
The cryptocurrency market is currently one of the most interesting areas for investment, attracting both experienced and casual investors. While it can offer high returns, it also poses significant risks due to its high volatility. In this context, artificial intelligence, particularly through deep learning and machine learning algorithms, has played a key role in developing applications that provide investment advice, aiming to maximize returns and reduce investment risks. This study proposes a system for forecasting the closing prices of ten of the leading cryptocurrencies currently available in the market, presented in a web application capable of making predictions ranging from one to four hours. To achieve this, different models using various machine learning and deep learning algorithms were analyzed and tested, including Recurrent Neural Networks, time series analysis algorithms such as ARIMA, and even some more conventional regression algorithms. For algorithm comparison, minute step Bitcoin price data over a 30-day period was used to forecast prices 60 minutes ahead. The model that showed the best performance was the GRU Neural Networks model, using all quote attributes, achieving a MAPE = 0.09%, MSE = 5954.89, RMSE = 77.17, and MAE = 60.20.
The cryptocurrency market is currently one of the most interesting areas for investment, attracting both experienced and casual investors. While it can offer high returns, it also poses significant risks due to its high volatility. In this context, artificial intelligence, particularly through deep learning and machine learning algorithms, has played a key role in developing applications that provide investment advice, aiming to maximize returns and reduce investment risks. This study proposes a system for forecasting the closing prices of ten of the leading cryptocurrencies currently available in the market, presented in a web application capable of making predictions ranging from one to four hours. To achieve this, different models using various machine learning and deep learning algorithms were analyzed and tested, including Recurrent Neural Networks, time series analysis algorithms such as ARIMA, and even some more conventional regression algorithms. For algorithm comparison, minute step Bitcoin price data over a 30-day period was used to forecast prices 60 minutes ahead. The model that showed the best performance was the GRU Neural Networks model, using all quote attributes, achieving a MAPE = 0.09%, MSE = 5954.89, RMSE = 77.17, and MAE = 60.20.
Description
Keywords
Forecasting Cryptocurrency Machine learning Time series Previsão Séries temporais Aprendizagem máquina Criptomoeda