Πρόβλεψη αστάθειας σε ανταλλακτήρια με μηχανική μάθηση
Financial markets volatility prediction using machine learning
View/
Keywords
Machine learning ; Decision trees ; LGBM ; Feature engineeringAbstract
Τhe purpose of this project is to study the behavior of machine learning applications in the environment of stock exchanges. In this regard, algorithms were developed and executed to forecast the volatility of stock exchange products. The implementation was developed n the context of a competition which was organized by Kaggle, an online community of data science and machine learning professionals. A series of experiments were performed using different techniques of implementation and their results were compared. Furthermore the performance gains of parallelization of CPU tasks as well as using a GPU were investigated. The organizer of the competition provided the data to be used both for the developed models’ training and as a mechanism for algorithms’ evaluation.