G. Amati, C. Carpineto and G. Romano. Query Difficulty, Robustness, and Selective Application of Query Expansion. In Advances in Information Retrieval, 26th European Conference on IR Research, ECIR 2004. Pages 127-137. Sunderland, UK. April, 2004.

Amati et. al. proposed the query difficulty, which is applied for the Query Performance Prediction.

The query difficulty measures the divergence of a term's distribution in the top-ranked documents from its distribution in the whole collection. The notion is also related with the DFRTermWeightingModels for QueryExpansion.

[WWW] Terrier employs the[WWW] Bo1, [WWW] Bo2 and [WWW] KL DFRTermWeightingModels.