Analysing a contingency table with Kohonen maps: A factorial correspondence analysis

Abstract

The Kohonen self-organizing algorithm is a powerful tool to achieve a categorization of vectorial stochastic data into classes. Many researchers use it to get a preliminary reduction of the data complexity in numerous application fields. They address some problems which are usually solved by means of statistical methods like Classification, or Principal Component Analysis. In this paper, we propose to extend this approach to another data analysis method: the simultaneous analysis of two qualitative variables which are crossed in a contingency table.