In this paper, an incremental supervised classifier called INC-ALVOT (incremental voting algorithm) is presented. This algorithm allows handle mixed data sets which do not keep in main memory. Besides, it allows that when the classification of a goal object was realize, new objects are incorporated in original database, carrying out a minimal operations for the classification of goal object with the expanded data set. Result obtained with the proposed algorithm and classical ALVOT algorithm on different real data sets is presented.
