Toward Domain-Driven Data Mining

Traditional data mining is a data-driven trial-an-error process. It stops at discovered pattern/rule, either views data mining as an autonomous process, or only analyzes the issues in an isolated and case-by-case manner. As a result, the knowledge discovered is not interesting and actionable to constrained business. However, in many real world data mining tasks, for instance financial data mining in capital markets are highly constraint-based and domain- oriented. This paper proposes a new methodology named domain-driven data mining (DDDM), aims to discovery interesting and actionable knowledge for real user needs, overcome the gap between academia and business. DDDM integrates domain knowledge, expert experience, user interestingness, rule action ability and data into mining system. In this paper, A few basic concepts and methodologies are introduced firstly, after that the architecture is proposed and working detail is addressed. Finally, we specify issues that are either not addressed or insufficiently suited yet.