The Research And Implementation Of XML Information Retrieval Based On Language Models

With the rapid development of Internet technology, the quantity of information is increasing exponentially on Internet. Therefore, to enable people to obtain useful information effectively from the massive amount of information, a variety of Internet information retrieval services came into being. However, the task of just retrieving documents cannot satisfy the information needs of people. People hope that the information, which people wanted, can be more accurate, and have smaller particle size. But, as a self-descriptive markup language, XML has solved the problem to a great extent. Therefore, XML information retrieval has become one of hot spots of information retrieval research.In the XML information retrieval, effective retrieval models and indexing mechanisms can improve the overall performance of retrieval systems. Currently, on retrieval model, basically it uses the traditional information retrieval model, but reseaches language model rarely for XML information retrieval. But, the advantage of language model in information retrieval has been proved in many researches. Therefore, research language model for XML information retrieval is very significant.This thesis presents a XML information retrieval model based on language model, it is mainly to research and improve the popular retrieval model in information retrieval-language model, and enable it to carry out the structure and content (CAS) retrieval of XML, so as to improve the performance of XML information retrieval. Secondly, for adapting the retrieval model, this thesis proposes an XML index based on the technology of inversion list, and the use of improved Dewey code to better storing and computing for the relationship between nodes, and the optimization of the code storage. Finally, this thesis implements the XML information retrieval system, and proves its effectiveness through experiments.