ASSESSING SEMANTIC INFORMATION OF VOLUNTEERED GEOGRAPHIC INFORMATION
by Gursimar Kaur
by Gursimar Kaur
The world of cartography and map making has changed dramatically with the advent of new technological innovations and emergence of the hand-held mobile devices. The features like web mapping and navigation using electronic maps have made the paper maps outdated which lead to introduction of new phenomenon where the volunteers, also known as private citizens collaborate to share geographical information using a popular project called as Volunteered Geographic Information (VGI). Inspired from Wikipedia, OpenStreetMap (OSM) is the most successful project of VGI used for web mapping. It allows its contributors the freedom of global participation to collaborate their local knowledge for open access to everyone. Due to the open tagging scheme, the contributors augment the noisy and ambiguous data as the users have the freedom to use either previously generated tag or define their own. Further, a strict specification model is not used to audit the quality of the contributed data. The aim of the study is to assess the semantic similarity of the tags used to name the geographical feature by taking help of various string searching algorithms. This study implemented the algorithms to measure semantic similarity score of data under observation by assessing the attributes of tags and further divided the results as acceptably similar or not depending on the desired threshold value. The assessment of positional accuracy of linear features depicting real world geographical representation was achieved using the technique involving the creation of a constant width buffer around a line when a circle of fixed distance (also named as epsilon band) is rolled along both sides of the line. The designed approach helped to achieve data completeness and analyse the level of correlation in the given attribute constraints. Comparing the features with the dataset of higher accuracy, the evaluation is not limited to OSM and can be generalized using any other database crowdsourced by volunteers. The developed algorithms gave its contribution to enhance the enormous potential of the ever-rich dataset by improving its quality and alleviating the semantic gap in geospatial information.