Enhancing the quality of place resources in geo-folksonomies

Users’ interaction and collaboration on Web 2.0 via social bookmarking applications have resulted in creating a new structure of user-generated data, denoted folksonomies, where users, Web resources and tags generated by users are linked together. Some of those applications focus on geographic maps. They allow users to create and annotate geographic places and as such generate geo-folksonomies with geographically referenced resources. Geo-folksonomies suffer from redundancy problem, where users create and tag multiple place resources that reference the same geographic place on the ground. These multiple disjointed references result in fragmented tag collections and limited opportunities for effective analysis and integration of data sets. This paper, (1) defines the quality problem of resources in a geo-folksonomy (2) describes methods for identifying and merging redundant place resources and hence reducing the uncertainty in a geo-folksonomy, and (3) describes the evaluation of the methods proposed on a realistic sample data set. The evaluation results demonstrate the potential value of the approach