Building Real-Time Ontology Based on Adaptive Filter for Multi-Domain Knowledge Organization

Multi-domain knowledge organization is an effective way of correlating cross-domain knowledge or intercommunicating between cross-domain knowledge systems. As a knowledge organization model, ontology is widely used in information and management systems. To organize multi-domain knowledge, ontologies in different domains correlate to each other directly or indirectly. Generally, matching and integrating ontologies of different domain into a large scale ontology is the common way of directly correlating, while building a top level ontology is the main method for indirectly correlating. As the scale of domain ontologies get larger and larger, both direct and indirect methods become more difficult and time-consuming. In order to improve the organization of multi-domain knowledge, this paper presents a novel ontology organization method to build real-time ontology by adaptive filter while user presenting requirements. Only the entities related to user requirements are integrated, while building a real-time ontology. Firstly, the method searches domain ontologies that are related to user requirements. Then sub-ontologies are extracted from search results by filter, and they are integrated into a new ontology under direction of filter, i.e. real-time ontology. Finally, four criteria are introduced to evaluate real-time ontology. The experiment results illuminate that real-time ontology perform excellently in accuracy, recall, correctness and especially time-consuming.