Toward knowledge-driven data mining
Warwick J. Graco, Tatiana Semenova, Eugene Dubossarsky · 2007
This paper highlights the need to move from a method-driven approach to a knowledge-driven approach to data mining. A number of issues are covered including the need to develop 'smart' data-mining algorithms which include expert mining and modelling knowledge, the need to use 'intelligent data' or data that contains both metadata and metaknowledge, the need to marry business knowledge with technical knowledge with data mining and the need to use intelligence and other qualitative analyses to determine where data-mining efforts should be focused.