Popularity-based top-k spatial-keyword preference query

Applications based on spatial data has become present in our daily lives. Spatial data can be used to represent objects such as roads, bus stops, restaurants and schools. Some of these objects maybe associated with a text (e.g. menu of a restaurant). The objects that have spatial location (latitude and longitude) and text are named spatio-textual objects. There are a large number of interesting spatio-textual queries that can be posed. For example, a tourist maybe interested in hotels (spatial objects) that have a lot of restaurants in its vicinity. In this paper, we propose a new query type named Popularity-based Top-k Spatial-keyword Preference Query. Giving a set of query keywords and a spatial vicinity of interest; this query returns the k best spatial objects of interest in terms of the number (popularity) of spatio-textual objects of reference in their vicinity that are textually relevant for the given query keywords. We propose new algorithms to process this query efficiently and evaluate the algorithms proposed in real datasets. The results show the efficiency of spatial-based algorithms for radius bellow 5km and the efficiency of algorithms with hybrid indexes (spatio-textual indexes) for the majority of the experiments.