Improving Cache Techniques performance for Advanced Operations on High Dimensional Data

ABSTRACT this paper, we present an economical tree based mostly categorization technique, referred to as iDistance, for K Nearest neighbor (KNN) search in a very high-dimensional mathematical space. IDistance partitions the information based on a spaceor data-partitioning strategy, and selects a point of reference for every partition. The data points in every partition are reworked into one dimensional price supported their similarity with regard to the indicator. This enables the points to be indexed using a tree structure and KNN search to be performed exploitation one-dimensional vary search. The selection of partition and reference points adapts the index structure to the information distribution. We conducted in depth experiments to judge the iDistance technique, and report results demonstrating its effectiveness. We additionally present a price model for iDistance KNN search, which can be exploited in query improvement. Keywords: High Dimensional Data, KNN Search, iDistance, LSH, Caching Framework