Discovering knowledge in a large organization through support vector machines

Abstract

Much of the information used by an organization is collected in the form of manuals, regulations, news etc. These are grouped into controlled documentary collections, which are normally digitized and accessible via a content management system. However, obtaining new knowledge from collected documents in an organization requires not only sound search and retrieval of information tools, but also the techniques to establish relationships, discover patterns and provide overall descriptions of the entire contents of the collection. This article explores the nature of knowledge and the role that occupy the documentary collections as a source of obtaining him knowledge. It also describes the collection of documents will be used along the exposure of this study and the techniques of processing information in order to obtain the desired results. This paper describes the use of computational methods, support vector machines in particular, in a large organisation for document classification. © 2008 Springer-Verlag Berlin Heidelberg.