Exploiting Indifference for Customization of Partial Order Skylines

摘要: Unlike numerical preferences, preferences on at-tribute values do not show an inherent total order, but skyline computation has to rely partial orderings explicitly stated by the user. In such orders many ob-ject are incomparable, hence skylines sizes become unpractical. However, Pareto semantics can be modified benefit from indifferences: result essentially reduced allowing user declare some incomparable as equally desirable. A major problem of adding equiva-lences is that they may in intransitivity aggregated order and thus efficient query proc-essing hampered. this paper we analyze how far strict relaxed while al-ways retaining transitivity induced ag-gregation. Extensive practical tests indeed about two mag-nitude when using maximum possible relaxation still guaranteeing consistency with all prefer-ences.