This paper proposes a partition-based uncertain high-dimensional indexing algorithm, called PU-Tree. In the PU-Tree, all (n)data objects are first grouped into some clusters by a k-Means clustering algorithm. Then each objectȁ9;s corresponding uncertain sphere is partitioned into several slices in terms of the zero-distance. Finally a unified key of each data object is computed by adopting multi-attribute encoding scheme, which are inserted by a B+-tree. Thus, given a query object, its probabilistic range search in high-dimensional spaces is transformed into the search in the single dimensional space with the aid of the PU-Tree. Extensive performance studies are conducted to evaluate the effectiveness and efficiency of the proposed scheme.
展开▼
