Automatic generation of query sessions using text segmentation

Abstract

We propose a generative model for automatic query refor- mulations from an initial query using the underlying subtopic structure of top ranked retrieved documents. We address two types of query re-formulations a) specification where the reformulated query expresses a more particular information need compared to the previous query; and b) generalization where the query is reformulated to retrieve more general information. To test our model we generate the two reformulation variants starting with topic titles from the TREC-8 ad hoc track as the initial queries. We use the average clarity score as a specificity measure to show that the specific and the generic query variants have a higher and lower average clarity score respectively. We also use manual judgements from multiple assessors to calculate the accuracy of the specicity and generality of the variants, and show that there exists a correlation between the relative change in the clarity scores and the manual judgements of specicity.