Supervised Learning for Automatic Classification of Documents using Self-Organizing Maps.
Dina Goren‐Bar, Tsvi Kuflik, Dror David Lev · 2000
Automatic Document Classification that corresponds with user-predefined classes is a challenging and widely researched area. Self-Organizing Maps (SOM) are unsupervised Artificial Neural Networks (ANN) which are mathematically characterized by transforming high-dimensional data into two-dimension representation, enabling automatic clustering of the input, while preserving higher order topology. A closely related algorithm is the Learning Vector Quantization (LVQ), which uses supervised learning to maximize correct data classification. This study presents the application of SOM and LVQ to automatic document classification, based on predefined set of clusters. A set of documents, manually clustered by domain expert was used. Experimental results show considerable success of automatic document clustering that matches manual clustering, with a slight preference for the LVQ.