Semantic Web-based Collaborative Knowledge Management System

On the era of knowledge and economy, people universally use the Internet to transfer or acquire all kinds of knowledge. Traditional method of acquiring knowledge is that people search the information on the Internet through search engines, which usually stores on Web servers. This method has some disadvantages. Firstly, there are some limitations of stretching bottleneck and one-point error in the system that belongs to lumped mechanism. Secondly, there is no search engine' s index that can over one-third of webs with index, which make them unable to collect interrelated information and return to the users' desktop. In addition, information on the Internet usually lacks of semantic and search engine universally execute the search by means of keyword match. Most of searches return a lot of irrelevant results and important results, which are relatively involved with original searching keywords, can' t be provided. Usually it is hard for people to find out their own needed information, which makes it easy for users to be immersed in the information ocean.With the successful application in Peer-to-Peer system, such as Napster, Freenet and Gnutella, it provides for data share with efficient and convenient solving scheme. Especially, its unstructured system is equipped with good characters, such as stretching, agility and Autonomy, which preferably solve the problems that universally consist in lumped-mechanism system. Especially for the search of P2P, it needn' t get through the Web server, also can' t be limited by the information document format and host equipment, and reach a deep that traditional cataloging search engine can' t compare with (including all the opening information resource on the Internet in theory). P2P provide for information search on the Internet with a completely new solving method.However, the present P2P system also has some limitations that only support weak semantic (even lack semantic) and coarse-grained (level of file) sharing, so searched by identified characters of files, it is hard for users to find their own needed information, and the content with many similar identified characters and irrelevant semantic is returned, which leads to much lavishness on band width and computer resource and greatly restrict to bring into play potential of P2P. Evidently, share short of semantic can' t either satisfy costumers' requests or use the systematical resource efficiently.The appearance of semantic web points out the direction for solving the problem that lacking of semantic make machines unable to automatically identify information. Semantic web is constructed on the basis of XML with self defining marker and RDF that expresses flexible data, and its destination is that applying with efficient standard and technology makes computers capable of more understanding the information on the Internet to deal with knowledge findings, data integration, and information navigation, etc. Its core layer is XML, RDF and ontology. In the semantic web, ontology is the foundation on solving data resource share and transfer in the semantic layer.Ontology is a formal specification of a shared conceptualization. It supply for the communication between people in a certain field and application system with a kind of common knowledge share. Ontology can assist users to realize knowledge share on a certain degree, and the reason why goal is put in very important position is that , firstly, ontology analyses and clarifies the structure of field knowledge, identify the essence and relation of field conception which lays the foundation for knowledge share; Secondly, ontology can put analysis on field in very important position, so that it avoids redundant analysis of field knowledge; Last, ontology adopts explicit semantic, unifieddefined terms and conceptualization to make knowledge share possible.Thus, the thesis adopts ontology to make a semantic enriched description about member data with resource character, and puts forward a knowledge frame with semantic Web. By realizing knowledge management of peer node efficiently, the system frame can either solve the problem in P2P system what it is difficult to look for because of semantic lacking, or supply a way to find out interested peer in P2P system, and realize the share and simulation of knowledge by the mutual coordination among the peer,,The following is the chief work and contribution in the thesis:Firstly, knowledge management involves with the management on two sides of both explicit knowledge and tacit knowledge, especially for the management of tacit knowledge it is a difficulty. The collaborative character on P2P system is beneficial for tacit knowledge to realize the transfer and share. The thesis discusses the character of knowledge management, and suggests using content management to realize the method of finding out the tacit knowledge on the ground of the Polanyi theory about proximal and tip of tacit knowledge. Then a knowledge frame based on semantic Web is put forward.Secondly, P2P has a lot of characters, widely used in peer computation, collaborative operation, search engine and file share etc. However, the technology of P2P is not mature, how to supply with all kinds of efficient serve on its system is also a solving important problem. The thesis analyses some method on how to realize the Web serving on the basis of P2P' s frame of application program.Thirdly, knowledge retrieve is the focus on the knowledge management. In order to execute the retrieve effectively, the thesis adopts ontology to make a semantic enriched description about member data with resource character. Then using the technology of ontology match realizes theknowledge retrieve.