Development of Trading Algorithms Supervising Tools Using Ensemble Methods and Neural Networks

The topic of the master thesis is “Development of Trading Algorithms Supervising Tools Using Ensemble Methods and Neural Networks”. The object is the digital currency market, the subject is a model for classifying trading transactions into profitable or unprofitable before they are executed. The purpose is to develop a machine learning model that is based on ensemble methods and neural networks for supervising trading strategy algorithms through solving the binary classification task for trading transactions (profitable / loss-making) before execution to reduce the number of loss-making transactions due to choosing wrong entry point. The relevance of the topic is justified by technological trends and the corresponding challenges in the financial sector as a whole and in the asset management industry in particular, related to the automation of the entire chain of asset management and trading processes from analysis of investment opportunities and opening a position to fixing profit or loss and settlements with all counterparties, as well as increase in the share of robotic trading on world trading floors. To achieve the goal of the work, in the first part, approaches to the use of various machine learning tools were studied in relation to the active management of a portfolio of financial instruments and active trading. In the second part of the work, various models of machine learning based on ensemble methods and neural networks for supervising trading transactions were developed and tested.