Study On Intelligent Information Retrieval Mechanism Of Web-based

With the continuous development of Internet technology, the network is increasingly becoming an important source of accessing information in work and live of the us. However, due to the opening and heterogeneous of the network, it is difficult for users to quickly find valuable information from the complicated mass of information. For the information on the WEB, the traditional approach is categoried artificially, but this approach has many disadvantages, such as demanding a lot of people, consuming much more money and classificated results'consistency is not so high. Therefore, how to classify the WEB pages reasonably and effectively has become an important research topic in the field of the information retrieval. This thesis studys about that.First of all, this thesis introduces the definition and development overview of information retrieval technology. Then it describes the three classical model of information retrieval technology. At the same time, it points out the characteristics and existing problems in the classic model of information retrieval. We choose probability model which is non-binary matching and less calculated amount as the key to the research, aiming at the deficiency of zero input or output excessively during accurate matching process about Boolean model. Basing on all of what we mentioned, we combines feature extraction of Chinese Word Segmentation and text sort algorithm together, constructs information retrieval model based on Web by modification of probability model, as well as applying this model in Web page of petroleum safety production. Meanwhile, we has determined threshold value and realized the classification of relevant document in petroleum safety production according to training given text.Experiments show that Web-based information retrieval model expressing through the WEB page of oil safe production, meets the query demand of these professional users in the oil safe production, at the same time it improves precision rate in this area of information retrieval.