Projects

Applying machine learning methods to support labour market policy making

Mgr. Miroslav Štefánik, PhD.

Members of project
Ing. Martina Halousková, Ing. Zuzana Košťálová, PhD., Mgr. Lukáš Lafférs, PhD., prof. Ing. Štefan Lyócsa, PhD., Ing. Veronika Remenárová

• Basic informations
• Characteristics

• Duration of the project: 07/2022 - 06/2026
  ID number: APVV-21-0360

• The project's central aim is to explore new options opened by the recent advancement of machine learning (ML) research. We consider two suitable data applications of ML techniques: i. impact evaluation of active labour market policies (ALMP) using administrative data of registered job seekers, ii online job search data analysis. Members of the research team have experience in both these data contexts, documented by publications, projects participation, and underlined by the interest of external project partners: Institute of Financial Policy (IFP) and an online job-advertisement company - Profesia. Under the first objective, we plan to apply ML techniques in the impact evaluation of Slovak ALMPs. For this purpose, in cooperation with IFP, we will develop a cleaned and documented evaluation dataset. The annually updated evaluation dataset will serve as a base in generating particular evaluation studies. Particular evaluation studies will benefit from a methodological activity focused on adjusting and applying new ML techniques for the Slovak data context. In this respect, we can build on the current cooperation of the research team. The project's second objective aims to explore the potential of big, online data in forecasting socio-economic development. More specifically, Profesia is interested in utilising the information on the total number of registered job vacancy postings in predicting labour market indicators, such as the unemployment rate or the total employment. We plan to propose the most favourable design of such a forecasting model. Moreover, we explore additional, freely available online data sources (e.g. Google search trends) and their potential in improving the prediction. Finally, having access to online job vacancy data, we aim to explore their further utilisation in social research. Building on an already rich list of studies using data collected by Profesia, we extend our interest more in the depth of the content of the online job vacancy postings.