Tailoring Taxonomies for Efficient Text Categorization and Expert Finding

AutorRobert Wetzker, Winfried Umbrath, Leonhard Hennig, Christian Bauckhage, Tansu Alpcan, Florian Metze
QuelleIEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology 

Content categorization by means of taxonomies is a powerful tool for information retrieval and search technologies. It improves the accessibility of data both for humans and machines and applications of automatic data characterization can be found all over the Web. While research on automatic categorization has mainly focused on the problem of classifier design, hardly any effort has been spent on determining how many categories are actually necessary for a successful classification. Given that modern retrieval systems are based on taxonomies of tens of thousands of categories, this question is important for it will help accelerating data access. In this paper we demonstrate empirically that already small subtrees of a taxonomy often enable reliable categorization. We compare several measures for the selection of category subtrees and investigate to what extent the reduction affects the classification quality. We consider applications in classical document categorization and in the upcoming area of expert finding and report corresponding results obtained from experiments with standard benchmark data.