Extending average precision to graded relevance judgments

• Authors: Stephen E. Robertson;Evangelos Kanoulas;Emine Yilmaz
• Venue: Proceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval
• Year: 2010

Evaluation metrics play a critical role both in the context of comparative evaluation of the performance of retrieval systems and in the context of learning-to-rank (LTR) as objective functions to be optimized. Many different evaluation metrics have been proposed in the IR literature, with average precision (AP) being the dominant one due a number of desirable properties it possesses. However, most of these measures, including average precision, do not incorporate graded relevance. In this work, we propose a new measure of retrieval effectiveness, the Graded Average Precision (GAP). GAP generalizes average precision to the case of multi-graded relevance and inherits all the desirable characteristics of AP: it has a nice probabilistic interpretation, it approximates the area under a graded precision-recall curve and it can be justified in terms of a simple but moderately plausible user model. We then evaluate GAP in terms of its informativeness and discriminative power. Finally, we show that GAP can reliably be used as an objective metric in learning to rank by illustrating that optimizing for GAP using SoftRank and LambdaRank leads to better performing ranking functions than the ones constructed by algorithms tuned to optimize for AP or NDCG even when using AP or NDCG as the test metrics.