How can computer science contribute to knowledge discovery?

Abstract

Knowledge discovery, that is, to analyze a given massive data set and derive or discover some knowledge from it, has been becoming a quite important subject in several fields including computer science. Good softwares have been demanded for various knowledge discovery tasks. For such softwares, we often need to develop efficient algorithms for handling huge data sets. Random sampling is one of the important algorithmic methods for processing huge data sets. In this paper, we explain some random sampling techniques for speeding up learning algorithms and making them applicable to large data sets [15,16,4,3]. We also show some algorithms obtained by using these techniques.