Keren and Ernest's 6.867 Machine Learning Final Project
The goal of the project is to use sentiment analysis of stock-related news articles to predict the direction of fluctuation for a given stock. We have chosen a group of 9 stocks, all of which belons to the NASDAQ index and are in the Technology/Computer/Internet sector. Together, we collected 1737 articles on over 20 stocks from Reuter's Key Development news. We cross the articles with the stock price data we collect on the day of and day after of a given news to determine whether a news article has positive or negative impact.
The report can be found in /writeup