Two-Phase Path Retrieval Method for Similar XML Document Retrieval.

The existing three-dimensional bitmap indexing techniques have a performance problem in clustering the similar documents because they cannot detect the similar paths. The existing path clustering method which is based on the path construction similarity is one approach to solve the problem above. This method, however, consumes too much time in measuring the similarities between the similar paths for clustering. This paper defines the expected path construction similarity and proposes two-phase path retrieval method which effectively clustering the paths using it. The proposed method solved the performance degrade problem in path clustering by filtering the paths to be measured using the expected path construction similarity. [ABSTRACT FROM AUTHOR]