Ontology based Knowledge Model Research

Ontology has been widely applied in biological science (large-scale gene ontology), publishing field (Dublin-core standard and knowledge classification ontology), medicine and health care field (cancer ontology), culture inheritance field (museum and artist ontology). As the fundamental element of semantic web technology architecture, ontology provides expressive framework for knowledge reusability, sharing, representation and reasoning. However, ontology resource is quite rare in social science in China, especially in ancient historical records researching domain, which has turned into a bottleneck of semantic web development and application. On the strength of a semantic system—Chinese Fundamental Historical Records Analysis System, although inherent complexity of Chinese natural language processing (NLP) tools, especially aiming to ancient Chinese characters, this paper tried to make some study on knowledge abstraction, representation and conceptualization based on semantic ontology techniques in specific domain -Chinese ancient historical records research. In view of historical records domain features analysis and ontology constructing standard, this paper proposes a knowledge model of Chinese ancient historical records field, which mainly includes domain knowledge features analysis, modeling architecture and applying results. The key part of this paper is the architecture of the knowledge model. There are three levels presented in the model framework, namely resources level, transformation level and representation level. The knowledge resources level is dealing with the scope of raw data, that is what and how much materials are going to be used according to system requirement. The other important task of this fundamental level is to integrate those heterogeneous data to a relatively consistent data structure for further processing. Secondly, the transformation level is the most complex and important part of entire architecture. This level is to turn the initially processed data into ontological conceptual knowledge systems. The difficulty of transformation process is building up rule sets and ontology instances by rule abstraction and knowledge extraction from data objective. Besides, the design and developing of semi-automatic ontology building toolkit would affect the efficiency and precision of ontology knowledge base. The third level is a cluster of various historical records ontologies, which could be called into diversified knowledge systems with different dynasty, people, event and granularity in the light of different application requirement due to scalable and extensible mechanism of OWL ontology. As a result of this knowledge model architecture, a large-scale semantic knowledge system of Chinese ancient historical records was established. There are 122 classes, 32 object properties, 28 datatype properties, 15 reasoning properties and 179,053 instances in the ontology cluster. The amount of processed Chinese Character reached 17,835,000. This model is extensive research of ontology theory and technology in new application domain, which might be beneficial for historical records studying approach, knowledge discovery and in-depth study of domain ontology.