질의범위측도를 이용한 k-최대근접질의 처리 방법에 관한 연구

A Study on the Processing of k-Nearest Neighbor Queries Using Range Measurement

Queries on the k-nearest neighbors (kNN) algorithm frequently occur in geographic information system (GIS), pattern recognition, spatial data analysis, and image processing to find the k spatial objects that are closest to a given query point. To increase the efficiency of processing these queries, it is necessary to accurately measure their range and create the buckets to be considered spatial localities. In this paper, we propose a strategy of bucket-creation to minimize the processing cost of kNN queries. We do this by experimenting with a kNN query-processing algorithm that can accommodate a measurement to estimate the ranges to be used by the nearest neighbor queries. Results demonstrated that our kNN query algorithm, based on the range-estimation method, improved the number of accessible disks with increasing values of k.