Career-path analysis using drifting Markov models (DMM) and self-organizing maps

Sébastien Massoni (1) , Madalina Olteanu (2) , Patrick Rousset (3)

1 CES - Centre d'économie de la Sorbonne
2 SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne)
3 CEREQ - Centre d'études et de recherches sur les qualifications

Analyzing school-to-work transitions is an important challenge for the specialists of the labor-market. The aim of this paper is to study the insertion of graduates and to identify the main career-paths typologies. We introduce a new methodology for clustering career-paths by combining statistical estimation of non-homogeneous Markov chains with self-organizing maps. The proposed methodology is tested on real-life data issued from the survey ''Generation 98'' elaborated by CEREQ, France (http://www.cereq.fr/)

Mots clés: Career paths; categorical data; drifting Markov model; self organizing maps

Domaines: Applications [stat.AP]; Machine Learning [stat.ML]; Economies et finances

Sébastien Massoni, Madalina Olteanu, Patrick Rousset. Career-path analysis using drifting Markov models (DMM) and self-organizing maps. MASHS, 2010, Lille, France. ⟨hal-00443530⟩