Machine learning analysis for improving the matching on the labour market
Project duration: 01.10.2019 to 30.09.2022
Abstract
Finding an appropriate job is crucial for a good match on the labour market. The placement process should be supported with statistical methods. The project will focus on several research questions related to the matching of vacancies and job seekers. First, given the information about the characteristics of certain jobs as well as jobseekers one can predict which category of job has the highest matching probability for a specific person. Furthermore, predictions can consider job quality for a jobseeker. For this purpose, one can look at wages or job durations for example. The predicted wage can serve as an indicator for job quality by comparison with the average wage for a certain job category. Beyond other data on structural characteristics, such analyses might use information on competences and (soft) skills. Hence, it could be determined whether and which skills are relevant for the realisation of matches.
Concerning the employer side, one main question in this context is finding out which persons with which characteristics are most likely to be placed on specific jobs.
Beyond, the concrete vacancies could be investigated. Given the characteristics of a certain open position, based on realised data one could analyse which persons proved to be best matched to such vacancies. Further, one could look at the job offer itself and check if there are some possible modifications that cause a sizeable improvement in matching.