Our own sentiment analysis here at Intelligent Trading Technology is that sentiment analysis’ star is rising. A few weeks back, I chimed in on the progress this segment of our market has made as evidenced by the sophistication and enthusiasm on show by both presenters and audience members at a recent Bloomberg-hosted seminar on machine learning.
Now, courtesy of Nikki Carpenter of Acuity Trading, who worked with us on a Dow Jones Newswires roadshow on sentiment analysis back in 2008, we’ve come across some research on the predictability of Accuity’s market sentiment data.
The paper is based on research undertaken by the Universitat Politècnica de Catalunya, and led by Professor Argimiro Arratia, author of Computational Finance: An Introductory Course with R. The research was conducted on Acuity’s eleven news-based public sentiment indices to identify which sentiment indicator or combination of indicators provided the most reliable forecast.
It’s pretty technical stuff, but we figured some of the news geeks out there would welcome a look under the hood of how some of this works.
I think you’ll find it fascinating.