| With the development of science&technology and economy, the electric data in units such as laboratory, news_publishing, enterprise become larger very fast, the information for people to choose also become very huge. These electric information usually don't be stored in the organized database,but stored in file system at the document type like txt, doc, html,xml, xls, ppt, pdf,eg. When people need particular information, they need to search the useful information from these large amount of documents.But now the full_text retrieval system seldomly satisfied the users on recall and precision when delealing chinese text, the result is in high degree be associated with the character of the chinese text.The research focus of the paper is chinese text segmentation. chinese text segmentation is very important in chinese information technology. At now time there are many kinds of chinese text segmentation algorithm. To improve the recall and precision and reduce divergent divided svncopation, the knowledge segmentation algorithm for chinese text is promoted in the paper. The algorithm is based on a knowledge system.The database full_text retrieval technology also get a fast development. All popular database products have the function of full_text retrieval. Because the search engine is embedded in the database system, the retrieval efficiency is high, and is easy to use and manage. The paper use the funtion oracle text embedded in oracle database system to establish a medrec information retrieve system, at the same time link with the knowledge segmentation algorithm, improve the recall and precision of retrieval system. Chapter1 demostrator the use and requirement ,reality, indicator the theory and the focus of the research. Chapter 2 introduce the theory and trends of full_text retrieval. Chapter 3 introduce tipical segmentation algorithm. Chapter 4 focus on the research and practice of knowledge segmentation algorithm. Chapter 5 introduce how to establish the medrec full_text retrieval system using the technology of oracle. |