Navigation in Databases Using Self-Organising Maps

Abstract

This chapter explores that knowledge must be structured and linked by associations. Taking into account the volume of accumulated data the process of data structuring must be highly automated. To this end self-organizing maps (SOM) are very well suited for converting data into knowledge. They are self-organizing and they provide clarity, making data perception easier for people. These features make self-organizing maps ideal navigation tool for huge collections of data. The goal of this chapter is to demonstrate the use of self-organizing maps in Internet and business applications. The chapter presents several applications of Kohonen maps for organizing business information—namely, for analysis of Russian banks, industrial companies, and the stock market. The chapter explains how to use self-organizing maps for navigation in document collections, including Internet applications. Finally, the chapter also discusses some possible extensions of presently used self-organizing maps.