Research And Implementation Of Semantic-based Key Technology For Information Retrieval

With the rapid development of network technology, information shows explosivegrowth. Information retrieval system has become an indispensable means to gaininformation in the Network world. However, the traditional information retrievaltechnology based on keywords cannot understand user’s real demand accurately, whichoften bring the massive trash information. Therefore, how to understand and extract theuser’s query intention in the semantic level, and then search out the most accurateinformation, become the most important field of information retrieval research.In this paper, a depth studies carried out for the dictionary-based Chinese sentencesimilarity calculation and its application in information retrieval, then propose a newalgorithm of similarity calculation vector-model which base on the syntax of Chinese.The algorithm based on the grammatical structure of Chinese sentences describes thesimilarity by analyzing the meaning and the syntactic features of sentences. Fullyconsidering the influence of the relationship between words and semantic contributionto the similarity of sentence, this algorithm design different weighting parameters fordifferent syntax elements which effectively improve the accuracy of the similaritycalculation.Then the application of the algorithm in semantic retrieval is given and anew sentence-level semantic retrieval model based on Intention Analysis is proposed.Finally, the algorithm was tested and compared with other semantic algorithmsin a closed environment. Experiments indicated that under the grammar informationcorrect premise, the algorithm which this paper proposed may obviously enhance theretrieval performance and achieves the good effect...