thanks for d ans.. Information Extraction is expressed as queries on parse treeTdatabase as the query languages such as XPath and XQuery is not suitable for extracting linguistic patterns and hence Parse tree Query language allows the users to define extraction patterns on grammatical structures such as  constituent trees and Linkages.

well reference to parse tree query language is in the attached paper [section 3.2].. In the paper there are two methods for generating the extraction queries in the presence and absence of training data automatically...

The methods are easy but still there is some confusion in that..I still need some clear idea on that..

Thanks