A Framework for Identifying Skylines over Incomplete Data

摘要: Skyline queries provide a flexible query operator that returns data items (skylines) which are not being dominated by other in all dimensions (attributes) of the database. Most existing skyline techniques determine skylines assuming values for every item available (complete). However, this assumption is always true particularly multidimensional database as some may be missing. The incompleteness leads to loss transitivity property technique and results into failure test dominance incomparable each other. Furthermore, influences negatively on process finding skylines, leading high overhead, due exhaustive pair wise comparisons between items. This paper proposed framework incomplete with aim avoiding issue cyclic deriving skylines. identifying consists four components, namely: Data Clustering Builder, Group Constructor Local Skylines Identifier, k-dom Generator, Incomplete Identifier. Including these processes has optimized reducing necessary number comparison through eliminating early possible before applying technique.