Predicting Stock Price Movement from Financial News Articles

The ability to predict stock market movement has been a source of interest for many researchers. While numerous scientific attempts have been made, no single method has yet been discovered to accurately predict stock price movement. Difficulty in prediction comesfrom the complexities associated with market dynamics where parameters are constantly shifting and not fully defined. One area of limited success in stock market prediction comes from textual data. Not all data begins first as quantitative data and ratios. Information from quarterly reports or breaking news stories, which can dramatically affect the share price of a security, begin as qualitative data and must be translated to nu-meric form before many computational systems can process it. This information lag could be capitalized on by applying computational methods to the textual data and forms the basis offinancial text mining. Most existing literature on financial text mining applies a representational technique to news articles where only certain terms are used, and weights are assigned to the terms based on the direction the stock price moves. Prediction then applies these weighted terms to a new article to determine a likely direction of movement. To their credit, these simpler forms of analyses have shown a weak but definite ability to predict price direction but not the price itself. However, using computational approaches to predict stock prices using financial data is not unique. In recent years, interest has increased in Quantitative funds, or Quants, that automatically sift through numeric financial data and issue stock recommendations. While these systems are based on proprietary technology, they do differ in the amount of trading control they have, ranging from simple stock recom- menders to trade executors. Using historical market data and complex mathematical models, these methods are constrained to make assessments within the scope of existing information. This weakness means that they are unable to react to unexpected events falling outside of historical norms. However, this disadvantage has not stopped fund managers at Federated, Janus, Schwab, and Vanguard from trusting billions of dollars of assets to the decisions of these computational systems.