From research at Stanford University, it was found that Twitter tweets were able to positively reflect the general public’s sentiments towards a given stock. Twitter allows all humans to post short updates within 140 characters. Through machine learning and sentiment analysis, machines are able to instantaneously make assertions about a stock based on public opinion.
Neural networks can be trained with public sentiments along with historical price data to more accurately predict future prices of stock. The paper was able to show a 75% correlation between their predictions and actual stock prices.
Technology is changing the game for the financial services sector. With the increased availability of compute power, it will push the entire world towards trading that is heavily dominated by technology instead of human traders. But, I believe this artificial intelligence needs to be augmented by humans.
Reference: https://pdfs.semanticscholar.org/4ecc/55e1c3ff1cee41f21e5b0a3b22c58d04c9d6.pdf