Collective spatial keyword querying

Abstract

With the proliferation of geo-positioning and geo-tagging, spatial web objects that possess both a geographical location and a textual description are gaining in prevalence, and spatial keyword queries that exploit both location and textual description are gaining in prominence. However, the queries studied so far generally focus on finding individual objects that each satisfy a query rather than finding groups of objects where the objects in a group collectively satisfy a query.We define the problem of retrieving a group of spatial web objects such that the group's keywords cover the query's keywords and such that objects are nearest to the query location and have the lowest inter-object distances. Specifically, we study two variants of this problem, both of which are NP-complete. We devise exact solutions as well as approximate solutions with provable approximation bounds to the problems. We present empirical studies that offer insight into the efficiency and accuracy of the solutions.

On Generalizing Collective Spatial Keyword Queries

“…Five different cost functions have been proposed for the CoSKQ problem, namely, cost Sum [3], cost M axM ax [3], cost M axM ax2 [17], cost M inM ax [2] and cost SumM ax [2]. For example, cost Sum (S) defines the cost to the summation of the distances from the query location to the objects in S, and cost M axM ax (S) defines the cost to a linear combination of the maximum distance between the query location and an object in S and the maximum pairwise distance among the objects in S. The definitions of the rest of cost functions would be introduced later.…”

“…For the CoSKQ problem with each of these existing cost functions, which was proved to be NP-hard, at least one solution (including an exact algorithm and an approximate algorithm) was developed, and these solutions usually differ from one another. For example, the exact algorithm for the CoSKQ problem with cost Sum is a dynamic programming algorithm [3], while that for the one with cost M axM ax is a branch-and-bound algorithm [3]. Usually, an existing algorithm for the CoSKQ problem with a particular cost function cannot be used to solve that with another cost function.…”

“…Among them, some [3], [17], [2] studied the collective spatial keyword queries (CoSKQ). Cao et al [3], [2] proposed four cost functions, namely cost Sum , cost M axM ax , cost M inM ax and cost SumM ax , and developed algorithms for the CoSKQ problem with the first three cost functions, leaving that with the fourth cost function, i.e., cost SumM ax , as future work. Besides, they studied two variations of CoSKQ, namely top-k CoSKQ and weighted CoSKQ, in [2].…”

“…A collective spatial keyword query (CoSKQ) [3] is defined as follows. Problem 1 (CoSKQ [3]). Given a query q with a location q.λ and a set of keywords q.ψ, the CoSKQ problem is to find a set S of objects such that S covers q.ψ, i.e., q.ψ ⊆ ∪ o∈S o.ψ, and the cost of S, denoted by cost(S), is minimized.…”

“…The above applications were captured by the so-called Collective Spatial Keyword Query (CoSKQ) [3], [17], [2] in the literature. Let O be a set of objects, where each object o ∈ O is associated with a spatial location, denoted by o.λ, and a set of keywords, denoted by o.ψ.…”