Similarity search on spatio-textual point sets
MetadataShow full item record
User-generated content on the Web increasingly has a geo tial dimension, opening new opportunities and challenges location-based services and location-based social netwo for mining and analyzing user behaviors and patterns. T applications of such analysis range from recommendat systems to geo-marketing. Motivated by these needs, que ing and analyzing spatio-textual data has received a lot attention over the last years. In this paper, we address problem of matching point sets based on the spatio-text objects they contain. This is highly relevant for users asso ated with geolocated photos and tweets. We formally de this problem as a Spatio-Textual Point-Set Join query, a we introduce its top-k variant. For the efficient treatm of such queries, we extend state-of-the-art algorithms spatio-textual joins of individual points to the case of po sets. Finally, we extensively evaluate the proposed me ods using large scale, real-world datasets from Flickr and Twitter.