Bachelor and Master Thesis

Boosting with Mutual Information

Boosting and Information Theory

Assume you can predict a target variable with 70% accuracy. If you combine 3 such predictions what is the accuracy then? The answer is, something between 78-90%. It depends on the mutual information. When you know which of the 70% correct predictions are in fact correct you can get up to 10% better boosting performances!

In this Bachelors or Masters thesis your goal is to find tighter bounds - something better than the 78-90% range - on the boosting performance by considering the mutual information. Your work will be on a theoretical level and requires motivation for math and programming in R. If you are interested, just drop me a line :)