Database Systems for Advanced Applications

Volume 6588 of the series Lecture Notes in Computer Science pp 134-148

wNeighbors: A Method for Finding k Nearest Neighbors in Weighted Regions

  • Chuanwen LiAffiliated withNortheastern University
  • , Yu GuAffiliated withNortheastern University
  • , Ge YuAffiliated withNortheastern University
  • , Fangfang LiAffiliated withNortheastern University

* Final gross prices may vary according to local VAT.

Get Access


As the fundamental application of Location Based Service (LBS), k nearest neighbors query has received dramatic attention. In this paper, for the first time, we study how to monitor the weighted k nearest neighbors(WkNN) in a novel weighted space to reflect more complex scenario. Different from traditional kNN approaches, the distances are measured according to a weighted Euclidean metric. The length of a path is defined to be the sum of its weighted subpaths, where a weighted subpath is relative to the weights of its passing regions. By dividing the plane into a set of Combination Regions, a data structure “Weighted Indexing Map”(WIM) is presented. The WIM keeps an index of the weighted length information. Based on WIM, we propose an efficient algorithm, called wNeighbors, for answering the WkNN query. The experimental results show that our WIM based WkNN processing algorithm are effective and efficient.


Nearest neighbor query Weighted Region kNN LBS Weighted Indexing Map