Ranking List Dispersion as a Query Performance Predictor

摘要

In this paper we introduce a novel approach for query performance prediction based on ranking list scores dispersion. Starting from the hypothesis that different score distributions appear for good and poor performance queries, we introduce a set of measures that capture these differences between both types of distributions. The use of measures based on standard deviation of ranking list scores, as a prediction value, shows a significant correlation degree in terms of average precision.