Domain Ontology Construction for Intelligent Anti-Telephone-Fraud Applications

Abstract

Domain ontology should reflect real data and support intelligent applications. However, existing methods for ontology construction focus on the conceptual architecture while ignore characteristics of domain data and application demands. To enhance the domain applicability of ontology, we propose a method of Data and Application Driven Ontology Construction (DaDoc), integrating data characteristics and application demands into the entire construction lifecycle. Our method includes three main phases: data and demands analysis, ontology construction and evaluation. Then, we apply this method to the anti-telephone-fraud field and construct the anti-fraud domain ontology based on cross-domain data. And this ontology represents anti-fraud domain knowledge and it can support intelligent anti-fraud applications, including semantic-based fraud identification and global situation analysis. We finally use quantitative indicators to evaluate the quality of anti-fraud domain ontology. Furthermore, this modeling practice validates the effectiveness of our method for ontology construction and this method can be applied to other domains.