Student's Perception of a Semantic Search Engine

This paper focuses on the question how students appreciate and value the possibility to query a multimedia knowledge base by entering complete questions instead of keywords. We aim at examining how far the result(s) of semantic queries are more appealing to students than those from a keyword-based search engine. In order to investigate this issue, two versions of the e-learning tool CHESt – an ontology-driven interactive expert system focusing on computer history – were tested in a secondary school. While the first version of CHESt implements a simple keyword search, the second version of CHESt carries out a semantic search. The aim was to assess whether students are satisfied with the number and the pertinence of the search results, and whether they generally appreciate the option to ‘communicate’ with the system by asking complete questions in natural language. The outcome of our investigation shows that students generally preferred to use the keyword instead of the semantic search function, independently from the judgment on the accuracy of the results yielded by the respective search engine. The results suggest that the pertinence of the results as judged by the students strongly depends on the familiarity of the users with both the formulation of questions and the domain of interest. Also the semantic search engine needs to be improved in order to extract more semantic information.