A signal-to-noise approach to score normalization

Score normalization is indispensable in distributed retrieval and fusion or meta-search where merging of result-lists is required. Distributional approaches to score normalization with reference to relevance, such as binary mixture models like the normal-exponential, suffer from lack of universality and troublesome parameter estimation especially under sparse relevance. We develop a new approach which tackles both problems by using aggregate score distributions without reference to relevance, and is suitable for uncooperative engines. The method is based on the assumption that scores produced by engines consist of a signal and a noise component which can both be approximated by submitting well-defined sets of artificial queries to each engine. We evaluate in a standard distributed retrieval testbed and show that the signal-to-noise approach yields better results than other distributional methods. As a significant by-product, we investigate query-length distributions.