Enabling effective location-based services for road networks using spatial mining

Abstract

A co-location pattern represents a subset of Boolean spatial attributes whose instances are located in a close geographic space. These patterns are important for location-based services. There are many methods for co-location pattern mining where the distance between the events in close geographic proximity is calculated using a straight-line distance called Euclidean distance. Since most of the real-time tasks are bounded to the road networks, the results computed using Euclidean distance is not appropriate. So to compute co-location patterns involving network we define a model where initially a network model is defined and the neighbourhood is obtained by using network distance. By comparing this approach with the previous Euclidean approach, the results obtained for co-location patterns on a road network are accurate. Our experimental results for synthetic and real data show that the proposed approach is efficient and accurate for identifying co-location patterns involving network entities.