PhD Predictive Analytics Projects! (NEW)

Share on LinkedInTweet about this on TwitterShare on Google+Share on FacebookShare on Reddit

genius at work title written with chalk on blackboard

Who’s Next?

In this new section, we are going to profile interesting Data Science and Predictive Analytics Projects by “up and coming” PhD and Masters students/recent grads.  The exposure may land them that first great job or simply start a compelling dialogue.  

All projects will be promoted here on the Data Science Report, externally to hiring managers, and relevant social media outlets, etc.  We will decide together on how to market the work so it gets noticed. 

PROJECT #1  BEATING THE SPORTSBOOK

Eric Chalmers

ERIC CHALMERS 
PhD Candidate (Expected Grad 2015)
Electrical / Biomedical Engineering

Using Machine Learning, Eric created a computer model in C# and Matlab which predicts the outcome of mixed martial-arts (MMA)/UFC fights and calculates how much to bet on each fight. It started as a for-fun side project – an experiment to see if a computer model could make predictions at least as accurately as human MMA/UFC fans.  He didn’t really expect it to succeed, but it represented a technical challenge he found interesting.

The easiest place to test the model against humans is the betting market, as the betting odds represent the collective predictions of thousands of MMA enthusiasts. Using a hypothetical 10,000 bankroll to start, the model’s predictions between January and today have a total return of 160%! Professional gamblers would call that amazing!

So we’re now following the model’s predictions through the next several months, to see if it keeps up this level of performance. View Project and See Predictions Here

 

__________________________________________________

PROJECT #2 “Your Project Here”

If you have an interesting project you’d like to highlight email me Ted O’Brien, Directer Data Science & Advanced Analytics Practice & Publisher of the Data Science Report Email: ted@starbridgepartners.com

Share on LinkedInTweet about this on TwitterShare on Google+Share on FacebookShare on Reddit