关键词: 本体, 变精度粗糙集, 概念格, 属性约简

Abstract:

The defects of existing ontology construction technology are analyzed in order to construct the ontology model of the ecommerce system. The model of ontology construction based on rough concept lattice is proposed for the sake of making up these defects by integrating the theory of variable precision rough set (VPRS) model and formal concept analysis. The method of βupper and lower distribution reduction is used to reduce formal context by improving the two algorithms of βvalues select and attribute reduction based on discernibility matrix in VPRS. Therefore, the paper proposes the reduction algorithm of concept lattices based on VPRS. After calculating the semantic similarity of concepts, the domain ontology model of ebusiness is built combined with knowledge of domain expert and original ontology model of the United Nations Standard Products and Services Classification Code by way of core ontology. Experiments show that the efficiency of building ontology based on rough concept lattice is higher than FCA.

Key words: ontology;variable precision rough set;concept lattice;attribute reduction