Textual Analysis of News for Stock Market Prediction

Stock market prediction constitutes an important factor in business. There are a large number of different mathematical models for predicting stock price movements. One of the alternative approaches is application of methods based on Natural Language Processing (NLP). Though NLP tasks are getting popular, they remain complex and voluminous. In the digital age almost, all information has been transferred to digital records that is good achievement. That is a good achievement. But on the other side because of the ease in creating new information, information search in Internet becomes complicated. This problem becomes more relevant, and scientists continue to search ways to structure a huge amount of information. There are many methods of traditional representations of words based on statistics. But these methods don't give representation about contexts and semantics of text document. In this paper, we will consider approaches that help to get semantics from news. To evaluate our methods, we will use them for predicting direction of S&P 500 Index. In other words, we will compare our approaches with a stock market prediction problem based on news.