What is ?² Distance

The aim of this paper is to present a typology of career paths in France drawn up with the Kohonen algorithm and its extension to a clustering method of life history analysis based on the use of Self Organizing Maps (SOMs). Several methods have previously been presented for transforming qualitative into quantitative information so as to be able to apply clustering algorithms such as SOMs based on the Euclidean distance. Our approach consists in performing quantitative encoding on labor market situation proximities across time. Using SOMs, the preservation of the topology also makes it possible to check whether this new method of encoding preserves the particularities of the life history according to our economic approach to careers. Lastly, this quantitative encoding preprocessing, which can be easily applied to analysis methods of life history, completes the set of methods extending the use of SOM to qualitative data.