Inexact discovery of knowledge with uncertainty

Inexact discovery of knowledge with uncertainty is one of the most important issues in the area of knowledge discovery and data mining. Traditional machine learning algorithms are mainly used for deriving exact rules. In the last decade, scientists have realized the importance of inexact discovery of knowledge with uncertainty. This paper explores the possible ways of extracting knowledge with uncertainty from observation. It presents two major categories of inexact knowledge discovery approaches: the probabilistic approaches and the possibility approaches. Existing problems of current approaches and possible solutions are also discussed