Algorithmic trading, mobile applications, usability testing, user movement analysis and prediction, network modeling and simulation, bio-inspired computing.
APVV-24-0446 funded by The Ministry of Education, Research, Development and Youth of the Slovak Republic.
Our research goal is to investigate how current machine learning approaches for time series prediction and classical financial chart analysis can be combined to create efficient trading models. We recognize that discretionary trading using classical approaches, such as financial chart patterns, candlestick formations, and divergences between price and indicators, is still widely applied. The primary objective is to analyze known chart patterns in novel contexts and identify new chart patterns, both of which we refer to as deep patterns. In our experience, this translates directly into a prediction problem, where a machine learning approach is used to estimate a (continuous floating-point) reward mapped to a financial chart. This reward reflects the potential price trajectory that may follow.
We further backtest our trained model several approaches for cross-validation - our simulation environment, QuantConnect and VectorBT. We also running Trading using Broker Alpaca on small test group of combination model-stock-parameters based on backtest results on Paper Account.
Mobile Computing Group