A study of the use of self-organising maps in information retrieval

Purpose We studied the applicability of self-organising maps for searching for information in a document collection. Design / methodology / approach – After conventional preprocessing, like transform into vector space, documents from a German document collection were trained for a neural network of Kohonen selforganising map type. Such an unsupervised network forms a document map from which relevant objects can be found according to queries. Findings Selforganising maps ordered documents to groups from which it was possible to find relevant targets. Research limitations / implications The number of documents used was moderate due to the limited number of documents associated to test topics. The training of self-organising maps entails rather long running times, which is their practical limitation. In future, our aim will be to build larger networks by