ES-Rank: evolution strategy learning to rank approach

Learning to Rank (LTR) is one of the current problems in Information Retrieval (IR) that attracts the attention from researchers. The LTR problem is mainly about ranking the retrieved documents for users in search engines, question answering and product recommendation systems. There are a number of LTR approaches from the areas of machine learning and computational intelligence. Most approaches have the limitation of being too slow or not being very effective. This paper investigates the application of evolutionary computation, specifically a (1+1) Evolutionary Strategy called ES-Rank, to tackle the LTR problem. Experimental results from comparing the proposed method to fourteen other approaches from the literature, show that ES-Rank achieves the overall best performance. Three datasets (MQ2007, MQ2008 and MSLR-WEB10K) from the LETOR benchmark collection and two performance metrics, Mean Average Precision (MAP) and Normalized Discounted Cumulative Gain (NDCG) at top-10 query-document pairs retrieved, were used in the experiments. The contribution of this paper is an effective and efficient method for the LTR problem.