Cryptocurrency price prediction using LSTM neural networks
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Tipo de documento: | Dissertação |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/10071/27575 |
Resumo: | The interest in cryptocurrencies is increasing among individuals and investors. Bitcoin is the leading existing cryptocurrency with the highest market capitalization. However, its high volatility aligns with political uncertainty making it very difficult to predict its value. Therefore, there is a need to create advanced models that use mathematical and statistical methods to reduce investment risk. This research aims to verify if long short-term memory (LSTM), and bidirectional long short-term memory (BiLSTM) neural networks, can be used with Savitzky–Golay filter to predict next-day bitcoin closing prices. We found evidence both networks can be used effectively to predict bitcoin prices. LSTM performed 4.49 mean absolute percentage error (MAPE) and BiLSTM 4.44 MAPE. We also found that using Savitzky– Golay filter and dropout regularization significantly improved the model’s prediction performance. |
id |
RCAP_05ffd249fb88d0b3372671d0f8eea6f6 |
---|---|
oai_identifier_str |
oai:repositorio.iscte-iul.pt:10071/27575 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Cryptocurrency price prediction using LSTM neural networksForecastingCryptocurrencySavitzky–GolayLSTMBiLSTMRede neural -- Neural networkPrevisãoCriptomoedaThe interest in cryptocurrencies is increasing among individuals and investors. Bitcoin is the leading existing cryptocurrency with the highest market capitalization. However, its high volatility aligns with political uncertainty making it very difficult to predict its value. Therefore, there is a need to create advanced models that use mathematical and statistical methods to reduce investment risk. This research aims to verify if long short-term memory (LSTM), and bidirectional long short-term memory (BiLSTM) neural networks, can be used with Savitzky–Golay filter to predict next-day bitcoin closing prices. We found evidence both networks can be used effectively to predict bitcoin prices. LSTM performed 4.49 mean absolute percentage error (MAPE) and BiLSTM 4.44 MAPE. We also found that using Savitzky– Golay filter and dropout regularization significantly improved the model’s prediction performance.O interesse em moedas digitais tem aumentado por parte de indivíduos e investidores. A bitcoin é a moeda digital com maior capitalização de mercado, no entanto, a sua alta volatilidade alinhada à incerteza política, torna muito difícil prever seu valor. Portanto, existe a necessidade de criar modelos avançados que utilizem métodos matemáticos e estatísticos para reduzir o risco de investimento. Este estudo tem como objetivo verificar se as redes neurais artificiais de memória longo curto prazo (LSTM) e redes bidirecionais de memória longo curto prazo (BiLSTM) podem ser usadas juntamente com o filtro Savitzky-Golay para prever os preços de fecho do dia seguinte da bitcoin. Os resultados mostraram que existe evidência que ambas as redes podem ser usadas de forma efetiva. LSTM obteve um erro percentual absoluto médio (MAPE) de 4.49 e BiLSTM um MAPE de 4,44. Também o uso do filtro Savitzky-Golay e regularização, melhora significativamente o desempenho de previsão dos modelos.2023-01-28T14:31:17Z2022-12-19T00:00:00Z2022-12-192022-10info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10071/27575TID:203180976engPereira, José Luís Almeidainfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2023-11-09T17:26:24Zoai:repositorio.iscte-iul.pt:10071/27575Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-19T22:11:49.570344Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Cryptocurrency price prediction using LSTM neural networks |
title |
Cryptocurrency price prediction using LSTM neural networks |
spellingShingle |
Cryptocurrency price prediction using LSTM neural networks Pereira, José Luís Almeida Forecasting Cryptocurrency Savitzky–Golay LSTM BiLSTM Rede neural -- Neural network Previsão Criptomoeda |
title_short |
Cryptocurrency price prediction using LSTM neural networks |
title_full |
Cryptocurrency price prediction using LSTM neural networks |
title_fullStr |
Cryptocurrency price prediction using LSTM neural networks |
title_full_unstemmed |
Cryptocurrency price prediction using LSTM neural networks |
title_sort |
Cryptocurrency price prediction using LSTM neural networks |
author |
Pereira, José Luís Almeida |
author_facet |
Pereira, José Luís Almeida |
author_role |
author |
dc.contributor.author.fl_str_mv |
Pereira, José Luís Almeida |
dc.subject.por.fl_str_mv |
Forecasting Cryptocurrency Savitzky–Golay LSTM BiLSTM Rede neural -- Neural network Previsão Criptomoeda |
topic |
Forecasting Cryptocurrency Savitzky–Golay LSTM BiLSTM Rede neural -- Neural network Previsão Criptomoeda |
description |
The interest in cryptocurrencies is increasing among individuals and investors. Bitcoin is the leading existing cryptocurrency with the highest market capitalization. However, its high volatility aligns with political uncertainty making it very difficult to predict its value. Therefore, there is a need to create advanced models that use mathematical and statistical methods to reduce investment risk. This research aims to verify if long short-term memory (LSTM), and bidirectional long short-term memory (BiLSTM) neural networks, can be used with Savitzky–Golay filter to predict next-day bitcoin closing prices. We found evidence both networks can be used effectively to predict bitcoin prices. LSTM performed 4.49 mean absolute percentage error (MAPE) and BiLSTM 4.44 MAPE. We also found that using Savitzky– Golay filter and dropout regularization significantly improved the model’s prediction performance. |
publishDate |
2022 |
dc.date.none.fl_str_mv |
2022-12-19T00:00:00Z 2022-12-19 2022-10 2023-01-28T14:31:17Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10071/27575 TID:203180976 |
url |
http://hdl.handle.net/10071/27575 |
identifier_str_mv |
TID:203180976 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799134672792322048 |