Estimating query performance using class predictions

We investigate using topic prediction data, as a summary of document content, to compute measures of search result quality. Unlike existing quality measures such as query clarity that require the entire content of the top-ranked results, class-based statistics can be computed efficiently online, because class information is compact enough to precompute and store in the index. In an empirical study we compare the performance of class-based statistics to their languagemodel counterparts for predicting two measures: query difficulty and expansion risk. Our findings suggest that using class predictions can offer comparable performance to full language models while reducing computation overhead.

• Type: Conference
• Year: 2009
• Where: SIGIR
• Authors: Kevyn Collins-Thompson, Paul N. Bennett
• Keywords: Class-based Statistics, Information Retrieval, Quality Measures, SIGIR 2009, Topic Prediction Data