Assessing the Quality of Search Process Models

摘要

Search processes are highly individual business processes reflecting the search behavior of users in search systems. The analysis of search processes is a promising instrument in order to improve customer journeys and experience. The quality of the analysis results depends on the underlying data, i.e., logs and the search process models. However, it is unclear what quality means with respect to search process logs and models. This paper defines search process models and revisits existing process model and log quality metrics. A metric for search process models is proposed that assesses their complexity and degree of common behavior. In order to compare metrics for search process models different logs and search processes are generated by using ontologies for user guidance during search process execution and for post processing of the logs. Based on an experiment with users in the tourism setting different logs and models are created and compared.