UCLA Master of Quantitative Economics Project for ECON 412
In this project, we would like to know if the text of news headlines can predict the direction of the market or a company? To answer such question, we plan to apply the method of Naïve Bayes to a real-world dataset with historic records of stock price variation and news text and try to predict the direction of stock price movement.
Kaggler Aaron7sun prepared a great dataset which includes the top 25 headlines from Reddit World News Channel and Dow Jones Industrial Average (DJIA) data from 2008/08/08 to 2016/07/01. There are 25 features that contain the top25 headlines text respectively. As for the only target, the label of 1 means the DJIA Adj Close value rose or stayed as the same; the label of 0 indicates the DJIA Adj Close value decreased.
More material will be uploaded through time. It is not surprising to see typos; please email me if you find any (anyiheng11@ucla.edu).