作者: Dennis Shasha , Tsong-Li Wang
DOI: 10.1145/96105.96111
关键词:
摘要: A scheme to answer best-match queries from a file containing collection of objects is described. query find the in that are closest (according some (dis)similarity measure) given target.Previous work [5, 331] suggests one can reduce number comparisons required achieve desired results using triangle inequality, starting with data structure for reflects precomputed intrafile distances. We generalize technique allow optimum use any set Some empirical presented which illustrate effectiveness our scheme, and its performance relative previous algorithms.