A novel tolerant skyline operation for decision making

Skyline operation is significantly important for Decision Support Systems (DSS) due to its ability to find a number of decision-maker (DM)-interested objects. However, an inherent weakness of conventional skyline queries is that the output size is hard to control by DMs. It actually includes two aspects. On one hand, the number of the returned skyline set might be too large to make the output meaningful. On the other hand, the skyline may not be informative enough or too concise to fulfill the DM’s interests. Current solutions for the first aspect aim to refine the computed skyline and find a representative skyline subset with a feasible size. But the second aspect still remains open. In order to tackle this problem, this paper attempts to extend conventional skyline and thus proposes a novel Tolerant Skyline Operation. We study algorithms for computing this novel tolerant skyline. The final experiments employ both real datasets and synthetic datasets for illustration of our approaches. The results indicate that the tolerant skyline is more effective and practical.