About
The Vague Query Responder was developed to address the limitations of traditional web search engines, which often failed to provide intuitively relevant results for vaguely-formed queries. It aimed to detect and exploit associations between query terms to reveal the underlying information needs in a user’s mind. The approach measured association strength between query terms using the CORDER method, identifying queries with low association strength as ‘vague queries.’ For such queries, WordNet was used to find related terms and extend the queries, relying on least common subsumers (LCS). The association strength refined these extended queries, which were further expanded using the Hyperspace Analogue to Language (HAL) model and the Information Flow (IF) method. Initial experiments on a corpus of 500 books from Amazon showed that this approach successfully identified relevant books for vague queries, even in cases where Google and Amazon’s book search tools failed.
