Νευρωνικά δίκτυα και εφαρμογές αυτών για την πρόβλεψη χρονοσειρών στο χρηματιστήριο
View/
Keywords
Νευρωνικά δίκτυα ; Πρόβλεψη χρονοσειρών ; ΧρηματιστήριοAbstract
The purpose of this thesis is the implementation and study of deep machine learning, and specifically Long Short-Term Memory (LSTM) recurrent neural networks, for short-term stock market forecasting. Deep machine learning represents a highly promising category of technical analysis, with LSTMs in particular proving very effective in the study of time series. Within the scope of this thesis, stock market forecasts are carried out using real stock prices (AAPL), and the factors influencing the performance of these forecasts are examined, with the ultimate goal of optimizing the results.