Top-k Typicality Queries on Uncertain Data

An uncertain object O can be modeled as a set of instances generated by an underlying random variable X. If there are a large number of instances in an uncertain object, how can we understand and analyze this object? An effective way is to find the most typical instances among all instances of the uncertain object. In Section 2.2.1, we applied the idea of typicality analysis from psychology and cognitive science to ranking uncertain data, and modeled typicality for instances in uncertain objects systematically.