A self-adaptive scope allocation scheme for labeling dynamic XML documents

Abstract

This paper proposes a self-adaptive scope allocation scheme for labeling dynamic XML documents. It is general, light-weight and can be built upon existing data retrieval mechanisms. Bayesian inference is used to compute the actual scope allocated for labeling a certain node based on both the prior information and the actual document. Through extensive experiments, we show that the proposed Bayesian allocation model can practically and significantly improve the performance of the conventional fixed scope allocation models. © Springer-Verlag Berlin Heidelberg 2004.