Predicting Query Performance via Classification

  • Kevyn Collins-Thompson ,
  • Paul Bennett

Proceedings of the 32nd Annual European Conference on Information Retrieval (ECIR 2010) |

Published by Springer

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 language-model counterparts for two performance-related tasks: predicting query difficulty and expansion risk. Our findings suggest that using class predictions can offer comparable performance to full language models while reducing computation overhead.