Data mining application issues in fraudulent tax declaration detection

Data mining technology has become a hot academic research area over the last ten years. Most data mining research concentrates on the model-building phase of the data mining process. In this paper, we focus on how to build data mining algorithm centered application system for common users. We cover the issues according to the phases in a data mining application system building process. These issues include communication with domain experts, choice of the core data mining algorithm, design guidelines of the data mining system architecture and incorporation of domain experts´ knowledge. To illustrate these issues, we present a case study about building a fraudulent tax declaration detection system using decision tree classification algorithm. The evaluation result of the system is also provided.