Fuzzy granular gravitational clustering algorithm

Given the nature of clustering algorithms of finding automatically, or semi-automatically, an unspecified number of clusters, much work has been done in this area. This paper will introduce a proposed gravitational model for finding clusters, the algorithm is based on the gravitational forces from Newton´s law of universal gravitation and the output clusters are then fuzzified. Two examples of datasets are compared, one synthetic and one of the Iris, are benchmarked against the fuzzy subtractive algorithm.