XML Vector Space Retrieval Of Automatic Granularity

This paper introduces XML information retrieval technology based on Vector Space Model and presents a retrieval method which can identify suitable information unit for retrieval in vector space automatically. It can adequately utilize the structure information of the XML documents to adjust the granularity of information unit dynamically based on actual query conditions and take the similarity between query and information units as the basis of the searches. Hence, the cost of computation is reduced efficiently and the performance of the whole information retrieval system is improved.