Metric All- K -Nearest-neighbor Search

Lu Chen,Yunjun Gao,Gang Chen,Haida Zhang

DOI: https://doi.org/10.1109/tkde.2015.2453954

IF: 9.235

2016-01-01

IEEE Transactions on Knowledge and Data Engineering

Abstract:An all-k-nearest-neighbor (AkNN) query finds from a given object set O, k nearest neighbors for each object in a specified query set Q. This operation is common in many applications such as GIS, data mining, and image analysis. Although it has received much attention in the Euclidean space, there is little prior work on the metric space. In this paper, we study the problem of AkNN retrieval in metric spaces, termed metric AkNN (MAkNN) search, and propose efficient algorithms for supporting MAkNN queries with arbitrary k value. Our methods utilize dynamic disk-based metric indexes, employ a series of pruning rules, take advantage of grouping, reuse, preprocessing and progressive pruning techniques, require no detailed representations of objects, and can be applied as long as the distance metric satisfies the triangle inequality. Extensive experiments using both real and synthetic data sets verify the efficiency of our proposed algorithms, compared with state-of-the-art Euclidean AkNN and MAkNN algorithms.