Counting aggregate classifiers

There are many methods to design classifiers for the supervised classification problem. In this paper, we study the problem of aggregating classifiers. We construct an algorithm to count the number of distinct aggregate classifiers. This leads to a new way of finding a best aggregate classifier. When there are only two classes, we explore the link between aggregating classifiers and n-bit boolean functions. Further, the sequence of the number of distinct aggregated classifiers appears to be new. / status: published