Retrieval of research-level mathematical information needs: A test collection and technical terminology experiment

Abstract

In this paper, we present a test col-lection for mathematical information re-trieval composed of real-life, research-level mathematical information needs. Topics and relevance judgements have been procured from the on-line collabo-ration website MathOvertlow by delegat-ing domain-specific decisions to experts on-line. With our test collection, we con-struct a baseline using Lucene's vector-space model implementation and conduct an experiment to investigate how prior ex-traction of technical terms from mathe-matical text can affect retrieval efficiency. We show that by boosting the impor-tance of technical terms, statistically sig-nificant improvements in retrieval perfor-mance can be obtained over the baseline.