Research On Method Of Forest Product Information Integration For Web Manufacturing Resource

Aiming at the characteristics of web manufacturing resources and user demands, the dissertation proposed a concept called "WEB information integration" and built the system architecture to integrate forest products resources on INTERNET. The main work of this dissertation dissertated in the followings:(1)Discovery of information resources. Through analysis different types of forest-related features of web pages features. A classification algorithm website based on key resource (TD-SVM) developed, which used to determine the type of websites. We find the number of forestry website that provide forest products information resources by web crawler program and TD-SVM algorithm on the Internet. The results of experiment for new website identification show that TD-SVM can be used in different area.(2) Information organization. Due to Text information of forest product information was disorder after information extraction. So it is need to organized. A developed Naive Bayes text classification system was developed for forest product knowledge to facilitate information application and management. In this part, The design and implementation of the text classification system are discussed for a well defined hierarchical forestry knowledge tree. Then, the key technologies such as feature selection, Naive Bayes algorithm, and classification are described in detail. The system can be used as an information infrastructure for the Chinese advanced "Digital forestry" system.(3) Information location. The traditional way of information retrieval is to use the keywords to express the information requirement. However, it is hard to faithfully express the user's retrieval purpose. So we proposed an information retrieval mechanism based on the forest product domain ontology. We use the domain ontology to comprehend and reasoning the original keywords inputted by the user. This chapter described a information location based on semantic retrieval.The dissertation proposed a model for WEB information integration, which discuss these aspects, included information acquisition, information organization and knowledge discovery. The model involves the use of the methods have similar needs in other areas of information integration practices has a certain reference value and significance.