Self-Organizing Maps of Large Document Collections

All applications presented in the previous chapters applied self-organizing maps to reducing quantitative, numeric data. This chapter shows how textual information can be treated in a similar way and how self-organizing maps can help in more effective retrieval of information than current search engines. The use of WEBSOM is a novel method for organizing collections of text documents into maps, for browsing and exploring links on the World Wide Web, or for organization of electronic messages or files. Timo Honkela and the team at the Neural Network Center at HUT provide several examples of the use of WEBSOM and many more are available on their website (http://nodulus.hut.fi/websomn/).