XML semantic search with natural language interface

With the application of XML document in digital library, science database and Internet, XML retrieval has been an important research field in information retrieval. One of the reasons for poor retrieval is that users´ need could not be described clearly. Query based on natural language could describe users´ need well. There are some researches on XML information retrieval based on natural language interface, but the technique is not suitable for retrieving XML document with Chinese. In this paper, we first propose an improved Chinese dependency parsing algorithm, and then give the XML structure and content mapping rules for a XML semantic search method with Chinese language interface. The Chinese dependency parsing algorithm improves the efficiency by reducing the word pair numbers to be judged of dependency relation by using the feature of Chinese. The experimental results show that the query terms can be extracted from Chinese query sentence efficiently by using the mapping rules we proposed.