Sentiment and Knowledge Based Algorithmic Trading with Deep Reinforcement Learning

Algorithmic trading, due to its inherent nature, is a difficult problem tackle; there are too many variables involved in the real-world which makes it almost impossible have reliable algorithms for automated stock trading. The lack of labelled data that considers physical and physiological factors dictate ups downs market, has hindered supervised learning attempts dependable predictions. To learn good policy we formulate an approach using reinforcement uses traditional time series price combines with news headline sentiments, while leveraging knowledge graphs exploiting about implicit relationships.