A Multi-View Voting Method to Combine Unsupervised Classifications

摘要

This article deals with the design of a new unsupervised classification results combination approach. This method is based on a new definition of the similarity between unsupervised classification and the notion of corresponding classes. This enables to use a new mechanism of unification to increase their efficiency. This mechanism is a multi-view voting approach which can compute a single result representing the whole of the results, taking into account the "opinion" of all the methods.