Clustering of Respondents for Developing a Recommender and Targeting System: The Case of Assigning Education, Training and Employment Services

The context of recommendation of education and employment services, for example in the so called "on-demand job markets", provides various marketing implications. Our study is based on empirical data from respondents to the questionnaires, the issuers of demand for training and employment, mainly students, as well as on answers of employers from the western part of Romania. We present an alternative analysis of the data without the use of socio-economic assumptions to guide results: using a simple evolutionary search enables explorative cluster formation. The comparison of clusters for student and employers opinions is based on the answer scores and illustrated by the emergence of distinct correlation structures for both types of respondents. At the end of our analysis we obtain the following information: the size of clusters, the comparison of cluster representatives, which turn out to be relative scores, which lead to cluster pairs with different degrees of opinion discrepancies between students and employers. High opinion discrepancies in cluster pairs inform which sub-population should be approached by educational re¬com¬men¬dation while low discrepancies would conduct to job placement recommendations or services.