Reverse spatial top-k keyword queries

Abstract We introduce the R everse S patial Top-k K eyword (RSK) query, which is defined as: given a query term q, an integer k and neighborhood size find all neighborhoods of that where q in top-k most frequent terms among social posts those . An obvious approach would be to partition dataset with uniform grid structure cell identify cells this keywords. However, answer incomplete since it only checks for are perfectly aligned grid. Furthermore, every (square) answer, we can define infinitely more result by minimally shifting square without including it. To address that, need contiguous regions any point region center satisfies query. propose algorithm efficiently RSK using index consisting augmented materialized lists frequencies. apply various optimizations drastically improve latency against baseline approaches. also provide theoretical model choose optimal minimize latency. further examine restricted version problem (RSKR) limits scope efficient approximate algorithms. Finally, how parallelism performance balancing workload smart load slicing technique. Extensive experimental evaluation proposed methods real Twitter datasets crime report datasets, shows efficiency our accuracy model.