Comparison of XReal and XDMA for Keyword Search in XML Databases: A Case Study

Abstract

More research works have been carried out in the area of keyword search on XML databases. This is mainly because of the fact that users do not require knowledge of database schema and query language to search XML databases based on keyword queries. In the recent past, XReal and XDMA are the most prominent keyword search approaches for XML databases. Both XReal and XDMA make use of the numerical facts and hierarchical structure of XML databases. Never the less, these approaches are not reliant on schema information of XML data. In XReal, term frequency and document frequency have been mainly utilized. However, XDMA employs dual indices and mutual summation with the use of the frequency of keyword matching tags and data. In this paper, we discuss briefly the very important aspects of these approaches. Also, we compare these two keyword search approaches based on their methodologies and experimental results. It is found that both XReal and XDMA are generally effective XML keyword search approaches. Moreover, XDMA is comparatively slower than XReal for larger XML datasets. Notably, XDMA has identified a new keyword ambiguity and also addressed all ambiguities.