Subscribe to DSC Newsletter

Predictive Analytics: Man versus Machine Competition

We would like to organize a competition where human beings compete against computers, to make predictions. The concept is simple: the brain, after all, is capable of doing raw predictions - weather predictions, stock market predictions, economic predictions, and so on. Even the tiny brain of some animals seem better than computers or the human brain, for short-term weather predictions. Human or animal brains, as well as computers, rely on (sensor) data to make predictions.

How would you set up a robust test to accurately measure who is consistently better at predictions? Let's say that we focus on predicting the stock market: how do you design a test to compare human beings with computers, both feeding on big data, with regard to making successful buy/sell transactions? Evidently, if we have a large number of participants designing stock trading strategies, some are going to perform very well just by chance. We want to reward people who succeed thanks to either sheer brain or computer power, but not thanks to chance.

We believe that the human brain, just like a computer or server cluster (cloud) equipped with SAS, R, Python, NoSQL database, Hadoop or other software, is an analytic machine. Indeed, I was once offered a well paid Wall Street job where no programming skill was required: only an analytic brain good with contrarian strategies (they were mostly hiring MIT and Princeton graduates, so I was a bit of an outlier).

Here we are looking for suggestions and participants for a human vs. machine predictive analytics competition. If you are interested in this project, please contact us at [email protected] 

Related articles

Views: 2365


You need to be a member of AnalyticBridge to add comments!

Join AnalyticBridge

Comment by Vincent Granville on November 4, 2013 at 9:14pm

Regarding the stock market predictions, I would not underestimate the predictive power of the human brain: in particular, it's ability to blend various sources of macroeconomic data, summarize it, and decide. Humans can be good at detecting bubbles and leverage them - which I did in the stock market, the housing market, and now a new bubble - as long as you have the right amount of greed and fear, not too much, not too little. I actually did better, without algorithms, than many hedge funds, between 1998 (when I started trading) and today. I've had long period of no activity, but that's part of my trading strategy. On a shorter time period, algorithms can do better. And for high frequency trading, you can't beat an algorithm.

On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service