4.17 BACKTESTING NEWS SENTIMENT SYSTEMS
The impact of a news item is usually very short-lived and most likely represents the perspective of the author(s) of the news item. Moreover, the impact of a news item can also be significant and may create major changes in the life of the stock of a publicly traded company.
Since not all news is created equal, there is questionable value regarding the use of a backtesting system that could evaluate the impact of the news sentiment of a news item months or even years after the news was published.
To get the most value out of a backtesting system, the news sentiment of a news item needs to be determined when the news becomes public and before the actual impact of the news item. Such a backtesting system indicates the ability of the news sentiment system to predict the impact before the stock movement rather than showing an ability to measure the sentiment effect after the move took place. The differences are significant and should not be ignored.
Get The Handbook of News Analytics in Finance now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
