Finding superior skyline points for multidimensional recommendation applications

摘要: In a typical Web recommendation system, objects are often described by many attributes. It also needs to serve users with diversified range of preferences. other words, it must be capable efficiently support high dimensional preference queries that allow the user explore data space effectively without imposing specific weightings for each dimension. The skyline query, which can produce set guaranteed contain all top ranked any linear attribute combination, has been proposed this type applications. However, suffers from problem known as `dimensionality curse' size query result grow exponentially number dimensions. Therefore, when dimensionality is high, large percentage become points. This makes such system less usable users. paper, we propose stronger called core adopts new quality measure vertical dominance return only an interesting subset traditional An efficient processing method find points using novel indexing structure Linked Multiple B'-trees (LMB). Our approach superior progressively need computing entire first.