News and events aware stock price forecasting technique

One of the most trending and important topic in contemporary times is stock price forecasting. The time series analysis is the most fundamental and still remains the most widely adopted technique. However, this approach is not accurate, since it does not incorporate the external factors like news, events, etc. which can affect the stock price. In this paper, the authors have proposed a “news and events aware” regression based stock price forecasting technique. The news and events information related to AAPL stock is extracted from Google trend and Reuters website. The historical stock price data of AAPL (ticker name for Apple Inc. stock) is taken from Quandl website to build an initial time series model. The magnitude of the news/events is calculated by Google trend index value whereas the nature of news/events is evaluated by performing sentiment analysis on each set of news data. The median method is used to remove the outliers in the news data. The regression model is developed based on a 5-year news and weekly stock price data. The proposed method shows a high correlation between the news/events and weekly changes in stock price. The resultsprove that theimplemented technique in this paper can potentially outperform the conventional time series analysis in stock price forecasting.