Actionable Combined Pattern Mining

摘要

Real world data mining applications have urgent requests for actionable knowledge satisfied the needs of decision-making. However, traditional data-centered data mining cannot deliver expected outcomes to business or industry. Therefore, domain driven data mining on top of traditional data-centered mining involves ubiquitous intelligence such as domain and expert's knowledge for actionable knowledge discovery is introduced. Toward domain driven data mining, at first, this article reviews present actionable combined pattern mining approaches such as combined association rule mining and class-bridge rule mining for actionable knowledge. Then in order to obtain more actionable and interesting knowledge satisfied genuine needs of decision-making fully, in-depth combined pattern mining and combined class-bridge pattern mining algorithm implemented on multi-domains combined pattern mining are proposed in this study. Finally, other issues such as affecting the actionability of patterns in business or industry environment are further provided to guide future research on actionable combined pattern mining.