Learning Classifiers via Data Summaries

N. Suematsu, T. Nakayasu, and A. Hayashi (Japan)

Keywords

Machine Learning, Data Mining, classification, data sum marization

Abstract

We propose a classifier learning method in which data sum maries are constructed and classifiers are learned from the summaries. This method is suitable for large databases, since we can specify memory space occupied by data sum maries and new instances can be inserted into the sum maries incrementally. We also show empirical results, which indicates that out method performs well in compari son to C4.5, even though it learns from summaries.

Important Links:



Go Back


IASTED
Rotating Call For Paper Image