View More View Less
  • 1 Centre National de la Recherche Scientifique (CNRS) Programme de Recherche Infométrie, Institut de l'Information Scientifique et Technique (INIST) 2 allée du Parc de Brabois 54514 Vandoeuvre les Nancy cedex (France) 2 allée du Parc de Brabois 54514 Vandoeuvre les Nancy cedex (France)
Restricted access

Abstract  

This paper describes the implementation of multivariate data analysis: NEURODOC applies the axial k-means method for automatic, non-hierarchical cluster analysis and a Principal Component Analysis (PCA) for representing the clusters on a map. We next introduce Artificial Neural Networks (ANNs) to extend NEURODOC into a neural platform for the cluster analysis and cartography of bibliographic data. The ANNs tested are: the Adaptive Resonance Theory (ART 1), a Multilayer Perceptron (MLP), and an associative network with unsupervised learning (KOHONEN). This platform is intended for quantitative analysis of information.