... eliminate the inversions, which are typical in human-built indexes. Inversion is a method used byprofessional indexers by which they break the order-ing of the words in each index entry, and list the ... preserved in their original ordering. For training and evaluation purposes, we used arandom split of the collection into 90% training and 10% test. This yields a training corpus of 259 docu-ments and ... to the back- of- the- book index, based on a set of linguistic and information theoretic features. We begin by iden-tifying the set of candidate index entries, followedby the construction of...