An Indexing Approach for Efficient Supporting of Continuous Spatial Approximate Keyword Queries

Abstract

Current social-network-based and location-based-service applications need to handle continuous spatial approximate keyword queries over geo-textual streaming data of high density. The state-of-the-art continuous spatial-keyword query indexing approaches generally do not apply for approximate keyword matching. Aiming to this problem, this paper first proposes an indexing approach for efficient supporting of continuous spatial approximate keyword queries by integrating min-wise signatures into AP-tree, namely AP-tree+. AP-tree+ can employ one-permutation min-wise hashing method to support approximate keyword matching with a lower maintenance cost. Towards providing a more efficient indexing approach, this paper has explored the feasibility of parallelizing AP-tree+ by employing Graphic Processing Unit (GPU). A min-wise parallel hashing algorithm is used to parallelize the approximate keyword matching of AP-tree+. The experimental results indicate that (1) AP-tree+ can reduce the space cost by 11.5% compared with MHR-tree, (2) AP-tree+ can hold a comparable recall and 5.64x query performance gain compared with MHR-tree while saving 22.9% maintenance cost on average, and (3) the GPU-aided AP-tree+ can averagely accelerate 5.76x faster than AP-tree+.