Goonatilake and Herath (2007) focus on the effect of news that surfaces throughout the day in the stock market. News stories about publicly traded companies were labeled positive, negative or neutral according to price changes in the company stock. Takahashi, Takahashi, Takahashi and Tsuda (2007) use Naive Bayes classifier for text measure stock price change before and after publishing news. They performe a morphological analysis and pattern matching to extract keywords then good, bad and neutral news clusters are created. The text categorization system perfom about 80% of classification accuracy. Liang (2006) use web stock news volumns analysis.