A density-based clustering algorithm for earthquake zoning

A possibility of applying the density-based clustering algorithm Rough-DBSCAN for earthquake zoning is considered in the paper. By using density-based clustering for earthquake zoning it is possible to recognize nonconvex shapes, what gives much more realistic results. Special attention is thereby paid to the problem of determining the corresponding value of the parameter ɛ in the algorithm. The size of the parameter ɛ significantly influences the recognizing number and configuration of earthquake zones. A method for selecting the parameter ɛ in the case of big data is also proposed. The method is applied to the problem of earthquake data zoning in a wider area of the Republic of Croatia.