Spatial Keyword Queries: Top k-Spatial Keyword Search (TOPK-SK)

Abstract

With advances in geo-positioning technologies and geo-location services, there area unit a chop-chop
growing quantity of spatiotextual objects collected in several applications like location primarily based services and
social networks, within which an object is delineate by its spacial location and a collection of keywords (terms).
Consequently, the study of spacial keyword search that explores each location and matter description of the objects has
attracted nice attention from the industrial organizations and analysis communities. Within the paper, we tend to study 2
basic issues within the spacial keyword queries: high k spacial keyword search (TOPK-SK), and batch high k spacial
keyword search (BTOPK-SK). Given a collection of spatio-textual objects, question a question location and a collection
of query keywords, the TOPK-SK retrieves the nighest k objects every of that contains all keywords within the question.
BTOPK-SK is that the instruction execution of sets of TOPK-SK queries[1]. Supported the inverted index and therefore
the linear quadtree, we tend to propose a completely unique index structure, known as inverted linear quadtree (ILQuadtree), that is rigorously designed to take advantage of each spacial and keyword primarily based pruning
techniques to effectively cut back the search area[2][3]. An economical algorithmic program is then developed to tackle
high k spacial keyword search. To any enhance the filtering capability of the signature of linear quadtree[4], we tend to
propose a partition primarily based methodology. Additionally, to alter BTOPK-SK, we tend to style a brand new
computing paradigm that partition the queries into teams supported each spacial proximity and therefore the matter
connectedness between queries[5][6]. We tend to show that the IL-Quadtree technique may with efficiency support
BTOPK-SK. Comprehensive experiments on real and artificial information clearly demonstrate the potency of our
strategies.