A Semantic Knowledge Fusion Method Based on Topic Maps

Knowledge fusion plays an important role in the integration of multiple, distributed, heterogeneous knowledge sources. This paper presents a semantic knowledge fusion framework based on topic maps. The fusion method is through measuring the semantic similarity between topic maps of knowledge object pairs. The similarity measure of topic maps consists of two steps: the syntax similarity of topic and the structural similarity of topic. The overall semantic similarity is computed by combining the two similarities with weight. The knowledge fusion flow and the algorithm that merge the topic maps are also proposed.