Research And Implement Of Noninteractive Literature-based Knowledge Discovery System

With the development of technological and the application of communication technology, information has exponential growth. Social division and the polarization of discipline led to fragmentation of knowledge.The explosing information and fragmentation of knowledge make it difficult to acquire knowledge, and it is more difficult to acquire knowledge that hidding in literatures which are no relation on the surface.Since the middle of last century, information workers have begun to study on fragmentation of knowledge.In1986,Don R.Swanson,the professor at the University of Chicago,broached the study of nointeractive literature-based knowledge discovery by a series of research.The study provides a beacon for digging hidden knowledge of the literature.Due to the development of noninteractive literature-based knowledge discovery system, we use computer to deal with complicated literature processing. The computer return results which provide the most likely prompt, that save lots of time and energy and promote the application of noninteractive literature-based knowledge discovery.Firstly, this paper sums up the exsiting noninterative literature-based knowledge discovery systems at home and abroad. By comparing those software systems, this paper makes a summary of those softwares desired to be.Secondly, this paper carries on studing noniterative literature-based knowledge discovery by a new set of ideas. Because the existing noninteractive literature-based knowledge discovery systems are mostly based on single language, medical literature and are mostly inmature and unpublished, this paper puts forward to study and design a noninteractive literature-based knowledge discovery system that used in English and Chinese, and no subject restriction. Chinese word segmentation is a hotspot in Chinese information process; there are lots of words that have no pratical meaning and have nothing to do with the study; natural language which used widely, freely and flexibly leads some difficult to information process and there are many words that are means widely and blurred, and can be replaced by other words. The goal of noninteractive literature-based knowledge discovery is to find middle words and target words, but there are too many words obtained by calculating. Therefore, this paper focuses on the solutions of Chiness word segmention, the filter of words, controls of synonym and the choice of middle words and target words.Thirdly, this paper introduces the design and realization of noninteractive literature-based knowledge discovery system. And the system is decomposed to modules. Then, following is the function description of each module, the realization of the system and brief introduction of the system.Finally, the noninteractive literature-based knowledge discovery system is used to evaluate its effect. In order to insure the validity of test results, the test results must be true and be persuasive. Rediscovering Swanson’s discovery, this paper aims to illustrate the effect of the noninteractive literature-based knowledge discovery system by comparing with the result of Swanson.Chinese literature data set originates from the domains of biology and medicine fo CNKI(China National Knowledge Infrasturcture) from the year of1979to2006. English literature data set originates from the PubMed database from1960to1980. By using the system, both in Chinese literature data set and English literature data set, the open discovery process and the close discovery process get almost all the middle words of magnesium and migraine, and all the middle words of fish oil and raynaud’s. However, it can’t discoved some middle words, because there are no literature about middle word and start word or target word of the data sets. Thus, the test results show that the system is good in noninteractive literature-based knowledge discovery.