Study Of K-Dominant Skyline Algorithms For Incomplete Data Stream

Being able to provide decision-making information for users, skyline query has become an important subject in the field of data mining research. Most of the data flow in real application has the characteristics of incompleteness, high dimension and ordering, so the transitive dominance between objects is abate, making it difficult to directly applied to the data flow with missing values. Incomplete data flow k-dominate skyline technology can solve the problem. But the existing incomplete data flow based k-dominate skyline query algorithms take too much time and space.Based on the detailed analysis of the characteristics and defects of the k-dominate skyline algorithm on incomplete data, two kinds of improved optimization algorithms are proposed:(1)A k-dominate DIKSkyline algorithm based on sliding window is proposed. The algorithm mainly uses the k-dominate ability of the objects and the possibility of being an k-dominate skyline object, then creates Indexs for this two factors, with the arrival of data objects, timely eliminates the old objects inside the sliding windows, so improves the efficiency of the algorithm;(2)A VPKSkyline algorithm is proposed. The algorithm introduces the concept of virtual point in the process of analysis, can effectively reduce the candidate skyline objects and the comparison times between different objects, significantly improve the performance of the algorithm. Finally, compared DIKSkyline and VPKSkyline relative with other algorithms through experiment and theoretical analysis, it is proved that they are feasible and effective.