Πρόβλεψη χρονοσειρών στο χρηματιστήριο με την χρήση σύγχρονων νευρωνικών δικτύων
View/ Open
Keywords
Νευρωνικά δίκτυα ; Χρονοσειρές ; Μετοχές ; ΧρηματιστήριοAbstract
The present work proposes a generalized methodology for forecasting the short-term movement of multiple stocks using a unified machine learning model trained on a subset of S&P 500 stocks. The methodology includes the collection and processing of historical data, the extraction of features and their statistical analysis, as well as the development and evaluation of an LSTM model with optimized parameters. The results highlight the ability of the proposed model to effectively capture short-term fluctuations in stock price trends.


