Research On Key Technology Of Semantice-Based Enterprise Knowledge Management System

Goal of knowledge management is that the right knowledge can be captured by person in right time, and then help person solve problem in practice. Knowledge Management System (KMS) refer to a class of information systems applied to managing organizational knowledge. Though many fruit were gained in research of KMS, they have some limitation: (1) Representation of data lack of semantic mechanism; (2) It is more difficult to get retrieve task-relevant knowledge due to lacking of interests and task information description of user. (3) They did not consider the identification of peer-groups with similar interest and task-needs; (4) and there is problem of system heterogeneity in the enterprise.Ontology and Web Service are regarded as basis of technology in the dissertation. Centered-user's need, realize recommending knowledge, sharing knowledge among enterprise members and cooperation work between workers is researched in the thesis. On the basic of theory knowledge of chapter 2, the main researches of the dissertation are listed below:First, Extracting ontology Oriented-relational databaseAt present, most of ontologies are built by the hands of experts, which have many disadvantages such as complexity of the project, strong depending on expert, too slow for building and so on. It is an effective method for ontology development and resuing the existing knowledge to acquire domain semantic information stored in relational database. Thus, a method of acquisition semantic information based on reverse engineering of the relational database and building domain ontology is presented. Groups of mapping rules are put forward in the method, which include mapping mapping tables, mapping columns, mapping datatypes, mapping constraints and data migration. According to some mapping rules, domain terms, relation between them and instrances are extracted from the relational shemes. Finally, domain ontology is built after experts refine it.Second, Semantic user modelBased on ontological theory and collect basic information, knowledge documents and task-relevant information need, semantic user model is built through extracting semantic concepts from ontogoly and learning data. According to the model, the system can recommend knowledge for enterprise memebers and we use genetic algorithm to evolve the model.Third, Virtual community based on Semantic user modelForm of multi-virtual communities based on spectral clustering algorithm and semantic user model is presented. Model of virtual community that consists of five elements is built. In the virtual community, sharing work documents is regard as mode of cooperation work. According to cooperation mode among members, three kinds of work documents sharing is presented, which are inner-transaction sharing, between-transaction and role sharing. In the aspect of knowledge sharing, we propose calculation similarity and publishing knowledge catalog about members in order to explicit knowledge. Community members can capture domain experts through search user profile to sharing implicit knowledge.Fourth, Prototype design of enterprise knowledge management systemArchitecture of prototype design of enterprise knowledge management system is built based on theory of chapter 3, chapter 4 and chapter 5. The architecture implemented through three tiers which is storage layer, knowledge service layer and user layer. We adopt Web Service and OWL-S technology to solve the problem of heterogenous system and service with semantic supporting mechanism. According to semantic user model and virtual community in store layer, Knowledge sharing and cooperation work are implemented in knowledge service layer. We do some validation jobs to work of dissertation under back ground of Sinoma (Su Zhou) Construction co., Ltd.