Purpose: A higher level of retrieval relevance along with pertinence is what information systems are required to provide if they are to gain more user satisfaction. Although in almost every IR system, general and non-topical terms are not considered to play a major role in indexing, the underlying assumption of current study stems from the fact that these terms could be effective in better searching by end-users. Thus this paper aims at proposing a new approach in this regard using a small thesaurus of general and non-topical terms which can be used along with queries (i.e., topical keywords). This would help making users’ queries closer to natural language. Method: In the first phase of this study, a set of 669 non-topical terms, which were identified through a previous study, was assumed as the test bed. Based on the main goal of the present paper, the list was analyzed and divided into major categories. As for the second phase, a preferred label representing each category was selected and assigned based on its “use warrant”; and to this end Google Trends was applied for determining the most-frequent general and non-topical terms among users’ web searches. At last the developed thesaurus was tested in searching and the retrieved results were evaluated in terms of relevance and pertinence. Results: The findings of this study show that, although there is rather a diverse range of general and non-topical terms appearing before or after topical keywords in Web documents, users are still using no or only a small fracture of them in their search queries. The findings also show that these terms would be of great assistance in providing more relevant results and a meaningful display of the results. This paper then proposes a conceptual model for applying such a thesaurus in searching the Web.

Keywords