Verifying the proximity and size hypothesis for self-organizing maps

The Kohonen Self-Organizing Mag (SOM) is an unsupervised learning technique for summarizing high-dimensional data so that similar inputs are, in general, mapped close to one another. When applied to textual data, SOM has been shown to be able to group together related concepts in a data collection and to present major topics within the collection with larger regions. This article presents research in which we sought to validate these properties of SOM, called the Proximity and Size Hypotheses, through a user evaluation study. Building upon our previous research in automatic concept generation and classification, we demonstrated that the Kohonen SOM was able to perform concept clustering effectively, based on its concept precision and recall7 scores as judged by human experts. We also demonstrated a positive relationship between the size of an SOM region and the number of documents contained in the region. We believe this research has established the Kohonen SOM algorithm as an intuitively appearing and promising neural-network-based textual classification technique for addressing part of the longstanding "information overload" problem.