Vote-boosting ensembles.

Vote-boosting is a sequential ensemble learning method in which theindividual classifiers are built on different weighted versions of the trainingdata. To build a new classifier, the weight of each training instance isdetermined in terms of the degree of disagreement among the current ensemblepredictions for that instance. For low class-label noise levels, especiallywhen simple base learners are used, emphasis should be made on instances forwhich the disagreement rate is high. When more flexible classifiers are usedand as the noise level increases, the emphasis on these uncertain instancesshould be reduced. In fact, at sufficiently high levels of class-label noise,the focus should be on instances on which the ensemble classifiers agree. Theoptimal type of emphasis can be automatically determined usingcross-validation. An extensive empirical analysis using the beta distributionas emphasis function illustrates that vote-boosting is an effective method togenerate ensembles that are both accurate and robust.