TMG-BoBI: Generating Back-of-the-Book Indexes with the Text-to-Matrix-Generator

A common editorial task when preparing scientific monographs, books, even long articles such as surveys, is to prepare a Back-of-the-Book index (BoBI). In this paper we study the automation of this task and propose a system for facilitating authors. We leverage text mining techniques and specifically Automated Keyword Extraction (AKE), the Text to Matrix Generator (TMG) which is a MATLAB toolbox for text mining, and the indexing tools of J4T_{E}X. From our experiments using publicly available datasets, our system appears to be capable of generating satisfactory indexes that can be used after only little user intervention.