Using Grouping and KNN Search Algorithm for High Length Catalo
B. Venkateswarulu, Y. Vinay Kumar · 2012
We propose a new group-adaptive space bound based on separating hyper plane boundaries of Verona groups to complement our group based catalog. This bound enables efficient spatial filtering, with a relatively small pre-processing storage overhead and is applicable to Euclidean and Mahalanobis similarity measures. Experiments in exact nearest-neighbour set retrieval, conducted on real data sets, show that our cataloguing method is scalable with data set size and data lengthily and outperforms several recently proposed cataloges. Consider approach for likeness search in interrelated, high-length data sets, which are derived within a grouping framework. They note that Catalog by “Vector Approximation” (VA-File), Which was proposed as a technique to combat the “irritation of Lengthily,” employs scalar quantization, and hence necessarily ignores dependencies across dimensions, which represents a source of sub-optimality? Grouping, on the other hand, exploits inter-length correlations and is thus a more compact representation of the data set. However, existing methods to prune irrelevant groups are based on bounding hyper-spheres and/or bounding rectangles, whose lack of tightness compromises their efficiency in exact nearest neighbour search. They propose a new groupadaptive space bound based on separating hyper-plane boundaries of Verona groups to complement their group based catalog.