Experimental Comparison of Six Fixed Classifier Fusion Rules

At present, fixed rules for classifier fusion are the most used and widely investigated ones, which, with no second- level training, compete with the more sophisticated rules. But, one problem with fixed rules is, that although they have good overall performance, it is not clear which one is good for a particular data set. In this paper, an experimental comparison of well-known six fixed rules (product, mean, maximum, minimum, median and majority voting) was done on some data sets of KDD’99, UCI and ELENA. The experimental results allow one draw some preliminary conclusions about comparative advantages of them.