• data mining
• cv
• automatic
• profiling
• text minig

The process of identifying job profiles from a set of curricula and selecting suitable candidates for a particular job offer is complex. The process is strictly time-consuming and expensive if we evaluate it in term of human resources. Different software tools were developed in order to simplify it for human recruiters. But with the advent of Industry 4.0, there has been a greater involvement from the enterprises in designing autonomous and data-driven approaches for recruiters. The aim of this thesis is to propose a scientific approach concerning keywords extraction and jobs profiling. Therefore we propose an approach to extrapolate a set of keywords that describes our job curricula and an approach to discover, time by time, the main career job profiles that populate the data set. We have also proposed a way to study profile evolution over the time. For the first objective, we exploit advanced Natural Language Processing and Text Mining techniques to automatically extract relevant keywords and build a semantic representation of curricula. For the second, we have exploited data mining techniques such as DBSCAN and OPTICS.