Web Based Instance Knowledge Item Auto Construction Method

With the rapid development of Internet, there is an explosive growth in the information of network . Compared with this, the knowledge of network is quite scarce. In this context, a variety of wiki knowledge websites based on the Web2.0 which provide a lot of valuable knowledge for users come to being and grow quickly. Internet information increases everyday, and the information which is with the rapid growth contains a lot of new knowledge. However, the existing knowledge library of Wikipedia is created by the majority of voluntary participation. This buliding approach is bound to face many issues such as that the lemmas are not updated timely. In order to solve these problems, this study developed a knowledge item auto construction system which filled the knowledge of lemma automatically and accurately and made up the existing shortage of Wikipedia.Based on the research of instance knowledge item auto construction method, we present a method which is effective for all fields of knowledge in the Internet. This system includes three main modules, which are information collection module, knowledge generation module and knowledge search module. This study covers the following areas:(1) Collection of information, gather all kinds of information which is relevant to the lemma from the Internet's massive information, including text, pictures, news, Bowen and video information.(2) Integration of knowledge, change the information gathered by the stage of collection of information into the knowledge which introduces one lemma by a lot of text processing.(3) Knowledge search, provide a user interface through which the users can not only access the knowledge already exists in the system, but also generate knowledge for new lemmas online.The system has been applied in the Haitian Yuan knowledge platform which is constructed by our laboratory successfully. Experimental results show that the proposed instance knowledge item auto construction method could meet the requirements of automatic construction of instance knowledge item, and the time complexity and the accuracy of knowledge all have reached the desired effect.