Key Technology Research On Topic Maps-based Knowledge Management

Topic Maps is a successful solution to organize concepts and the relationship among the concepts, which consists of 3 basic elements: topic, occurrence and association. It is useful for Topic Maps to define any complicated knowledge structure, organize the knowledge structure to information resource attached to metadata, and increase the speed for creating, maintaining and exchanging different metadata.Knowledge management field is researched on the basis of Topic Maps technology in the dissertation. The main creation points of this dissertation are summarized as follows:Firstly, the idea of metadata representation in the semantic web using Topic Maps instead of RDF (Resource Description Framework) technology is presented. The advantage of building a framework prototype for the semantic web based on Topic Maps is given according to comparing the characteristics between XTM and RDF on element, grammar, semantic and associated description ability.Secondly, integration method for Topic Maps-based heterogeneous knowledge is presented. The internal mechanism of the combination of Topic Maps and heterogeneous knowledge integration is discussed. The integration framework mainly made up of Topic Maps dictionary, system connector and heterogeneous knowledge integrator is designed, and the target of heterogeneous knowledge integration is realized. For the core part of the framework: Topic Maps dictionary, which is stored in the format of XTM, SCTM (Schema and Constraint for Topic Maps) specification and its processor are developed, the semantic validity of the Topic Maps dictionary is further checked, and the reliability of the framework is finally enhanced.Thirdly, design method for Topic Maps-based knowledge warehouse is presented. The logical framework of the knowledge warehouse is constructed by the orthogonal matrix made up of MDOM (Multi-Dimensional Ontology Model) and TM (Temporal Model). MDOM is represented by root-ontology and sub-ontology according to domain. Both ontologies include elements of topic class, association class and role class. The internal relationship of knowledge is represented through complementary ontologies. The instances are related to the knowledge warehouse to accumulate its volume through the element of occurrence class. Temporal serialization description for the elements in the knowledge warehouse is realized by Temporal Model, which is based on VTL (Valid Time Interval). Temporal serialization is important to temporal deduction, time-based constraints management and the retrieval for the history snapshot of the knowledge warehouse.Fourthly, TOM (Topic & Occurrence-oriented Merging) algorithm is presented to maintain the Topic Maps-based knowledge warehouse. TOM algorithm supports the merging from knowledge outside to the knowledge warehouse in the distributed circumstance by means of evaluating content similarity of Topic Maps.Fifthly, TMKMS (Topic Maps-oriented Knowledge Management System) prototype is presented. TMKMS solves the integration and share for enterprise resource through integration method for heterogeneous knowledge, and supports knowledge resource management for intellectual analysis through researching design method and maintain algorithm for knowledge warehouse.