This paper proposes a two-step methodology (a density-based clustering algorithm in the first step and SVMs (support vector machines) in the second step) to deal with GPS data without speed or acceleration features which are usually used as key attributes in the rule-based methodologies. ...