A clustering approach for identifying approachable locations using terrestrial surface transport

Detecting useful patterns from a given data by applying clustering algorithm has many practical applications. In order to perform the task of clustering identifying a set of good exemplars is a challanging job. Success of clustering greatly depends on the initial set of exemplar chosen as representatives. The paper proposes the use of Manhattan distance for identifying high quality exemplars that can act as an initial set of exemplars followed by iteratively refining them on the basis of resemblance between the different data points. The proposed algorithm has been efficiently implemented for identifying the important cities that are easily accessible from the other cities belonging to the same cluster.