Optimizing the search query for a document search engine

Use case: Swisslex

Optimizing the
search query for a
document search engine

Requirements

The provider of a legal information platform, containing collections of judgments, specialist journals, legal commentaries and works from specialized literature, wanted to improve the search results for the users of its search engine. Since the platform provides documents in up to 4 languages, multilingualism also played an important part in the search. To this end, the provider commissioned the ZHAW (Zurich University of Applied Sciences) to create a pilot project. The ZHAW then brought us on board to assist with consulting and implementation.

Solution

Although the university had considered creating an ontology, we were able to convince it that it would make sense to create a SKOS-based thesaurus instead, which would also enable it to optimize the keyword-based full-text search. The reason for this was that the competency questions required for creating an ontology to map specific properties were too imprecise and general. Therefore, an ontology, which would have been more elaborate to create, would not have led to better search results. Our tasks consisted of providing the ZHAW with the necessary infrastructure (VocBench server) as well as supporting them with personnel in obtaining part of the term candidates by term extraction and creating the pilot thesaurus.

Please also read the article published in the specialist journal edition(German).