Resumen de Unified Adaptive Resonance Theory Neural Networks for Numeric and Text data

Juthamas Choruengwiwat, Worapoj Kreesuradej

• This paper proposes a new neural network, called a unified adaptive resonance theory neural network, for clustering objects whose feature values may combine numeric and text data. The proposed model is based on the concepts of fuzzy ART neural networks and similarity measure of symbolic objects. The proposed model works directly on textual and numerical information without mapping objects onto some representations that have quantitative features. The inputs of the proposed neural network can directly receive both a textual value and a numerical value without mapping the text value into a numerical value. The preliminary experimental results show the unified adaptive resonance theory neural network that can correctly cluster data.