A novel Tolerant Skyline Operator for decision support

Skyline operator is significantly important for Decision oriented Data Analysis (DDA) due to its capability of finding a number of user-interested objects. However, an inherent weakness of conventional skyline queries is that the output size is hard to be controlled by users. It actually includes two aspects. On one hand, the number of returned skyline set might be too large to make the output meaningless. On the other hand, the skyline may be too concise to fulfill user'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 for the second aspect, it still remains open. In order to tackle this problem, this paper attempts to extend conventional skyline and thus proposes a novel Tolerant Skyline Operator. We also study algorithms for computing the tolerant skyline. The final experiments use real datasets for illustration of our methods. The results indicate that the tolerant skyline is more effective and practical.