In this blog series, international experts present their work on predicting the demand for skills using online job advertisements (OJAs). They explain whether and to what extent skills can currently be predicted by using OJAs and address opportunities and challenges. One of these experts is Elisa Margareth Sibarani from Leibniz University Hannover.

Job advert-derived time series for high-demand skillset discovery and prediction

At the 5th OJA Forum of the Bertelsmann Stiftung and the Federal Institute for Vocational Education and Training, Elisa Margareth Sibarani presented an approach for a job advertisement-derived time series for discovering and predicting high-demand skillsets using the SCODIS framework. SCODIS utilizes a graph-based method to represent skills in demand in the labor market, and specific critical indices (centrality, density) and cluster categories (isolated, secondary, principal, crossroads) that facilitate our ability to compute and reason about novel strategical-oriented observations. A critical next step is to extend our framework for comparing clusters over time. In this context, the performance of series generation is intended to be enhanced by leveraging the knowledge graph for semantic similarity, which currently relies only on the contents from each cluster.

 

Weitere Beiträge zum „5th OJA Forum“:

Johannes Müller – How can online job advertisements help to predict the demand for skills?

Kasper Kok – How can online job advertisements help to predict the demand for skills?

Fabian Stephany – How can online job advertisements help to predict the demand for skills?

Wyatt Clarke – How can online job advertisements help to predict the demand for skills?

Elisa Margareth Sibarani – How can online job advertisements help to predict the demand for skills?