The paper introduces a novel clustering method called Support Vector Clustering (SVC), which leverages the approach of support vector machines (SVMs). Data points are mapped to a high-dimensional feature space using a Gaussian kernel, where the minimal enclosing sphere is searched. This sphere, when mapped back to the data space, forms contours that enclose clusters. The width of the Gaussian kernel controls the scale of data probing, while the soft margin constant helps handle outliers and overlapping clusters. The algorithm identifies clusters by varying these parameters, maintaining a minimal number of support vectors to ensure smooth cluster boundaries. The method is demonstrated on several datasets, showing its effectiveness in handling noisy data and overlapping clusters. The complexity of the algorithm is analyzed, and its advantages over other clustering methods are discussed, particularly in generating cluster boundaries of arbitrary shape and handling high-dimensional data efficiently.The paper introduces a novel clustering method called Support Vector Clustering (SVC), which leverages the approach of support vector machines (SVMs). Data points are mapped to a high-dimensional feature space using a Gaussian kernel, where the minimal enclosing sphere is searched. This sphere, when mapped back to the data space, forms contours that enclose clusters. The width of the Gaussian kernel controls the scale of data probing, while the soft margin constant helps handle outliers and overlapping clusters. The algorithm identifies clusters by varying these parameters, maintaining a minimal number of support vectors to ensure smooth cluster boundaries. The method is demonstrated on several datasets, showing its effectiveness in handling noisy data and overlapping clusters. The complexity of the algorithm is analyzed, and its advantages over other clustering methods are discussed, particularly in generating cluster boundaries of arbitrary shape and handling high-dimensional data efficiently.