Skyline ranking for uncertain data with maybe confidence

Skyline queries have been actively studied lately as they can effectively identify interesting candidate objects with low formulation overhead. In particular, this paper studies supporting skyline queries for the uncertain data with "maybe" uncertainty, e.g., automatically extracted data. Prior skyline works on uncertain data assumes that every possible value for an uncertain object can be exhaustively enumerated (i.e., "alternatives" uncertainty) which is not applicable in many extraction scenarios. We develop fast algorithms that outperform the baseline approach by orders of magnitude and validate them over extensive evaluations.