Decision support system for predicting stock prices based on sentiments in social media

This paper proposes a theoretical model of a decision support system for predicting stock prices based on data from social media. The model corresponds to the concept of multi-agent information system, which is characterized by parallel work of separate intelligent agents. In contrast to other projects in this area, the new model uses neither technical analysis approaches nor econometric tools, but machine-learning algorithms that can connect social network content sentiments and stock market movements. The practical realization of this model can be used for commercial purposes.