Efficient approximate retrieval over multiple XML documents based on XPath-full-text query processing

In this paper, we propose an approach for efficient approximate retrieval based on the evaluation of XPath-full-text queries. We build the global structure index for all XML documents and then integrate the indexes using a numbering technique, so that the processing on structures is accelerated greatly. To process full-text queries on XML contents efficiently, a bloom filter is adopted for fast content retrieval. Experimental results show that our proposed methods are efficient and practical.