• automatic extraction
• database
• green economy
• job title
• job visualizer
• occupations
• text mining
• update
• work

Nowadays, the job market is continuously evolving, and elementary professional figures are progressively diversifying into more specialised profiles in numerous application fields. To categorize every occupation and its related skills and technologies, worldwide job databases have been developed. These frameworks share common difficulties concerning data completeness and upgrades, including manual and sporadic updates. By exploiting text mining techniques, this thesis focused on the development of an automatic method for the identification and extraction of job titles from text sources. Firstly, a set of clues was defined and collected to uniquely allow the identification of job titles within any text files. The collection of these markers was performed by analysing scientific papers, occupational databases, online dictionaries, and multimedia vocabularies. In a second step, the identified set of clues was utilized to automatically extract compound job titles. This extraction was achieved by generating a list of rules able to recognize any kind of complement that specifies the application area of each base job title.
To demonstrate its broad effectiveness, the implemented algorithm was tested in the meaningful topic of the green economy. Lastly, a demo-web application was created and published online to allow the user to assess the algorithm functionalities.
Overall, we believe this work could constitute a useful approach for worldwide database authorities, international recruitment companies, and institutions involved in professional formation and training courses.