Verifying the proximity hypothesis for self-organizing maps

The Kohonen Self-Organizing Map (SOM) is an unsupervised learning technique for summarizing high-dimensional data. When applied to textual data, SOM has been shown to be able to group together related concepts in a data collection. This article presents research in which we sought to validate this property of SOM, called the Proximity Hypothesis. We demonstrated that the Kohonen SOM was able to perform concept clustering effectively, based on its concept precision and recall scores judged by human experts. We believe this research has established the Kohonen SOM algorithm a promising textual classification technique for addressing the long-standing `information overload' problem.