This paper aims to conduct a comprehensive study on the listwise approach to learning to rank. The listwise approach learns a ranking function by taking individual lists as instances and minimizing a