Cryptocurrency price prediction using LSTM neural networks
Πρόβλεψη τιμής κρυπτονομίσματος με LSTM νευρωνικά δίκτυα
View/
Keywords
Πρόβλεψη ; Κρυπτονομίσματα ; LSTM ; Νευρωνικά δίκτυα ; Προεπεξεργασία δεδομένωνAbstract
Cryptocurrencies have been growing in interest, among all sorts of people and mostly investors. Above all cryptocurrencies, Bitcoin shows the highest amount of price volatility, driven by numerous political and economic reasons that make the trajectory hard to predict. Therefore, this paper tries to find out the efficiency of an LSTM neural network pre-processed with some data in forecasting the next day closing prices of Bitcoin. From our results, this LSTM model was able to efficiently predict the prices of Bitcoin with a mean absolute percentage error of 0.07. Moreover, standardization and normalization were applied, elevating its predictive power for the model enormously.
Department
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
53Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Όχι Παράγωγα Έργα 3.0 Ελλάδα