A new interpoint distance-based clustering algorithm using kernel density estimation

A novel nonparametric clustering algorithm is proposed using the interpoint distances between members of data to reveal inherent structure existing in given set data, where we apply classical univariate kernel density estimation method estimate around a member. Our simple its formation and easy resulting well-defined clusters. The starts with objective selection initial cluster representative always converges independently this choice. finds number clusters itself can be used irrespective nature underlying by an appropriate distance measure. analysis carried out any dimensional space viability high-dimensional use. distributions or their are not required known due design our procedure, except assumption that possess function. Data study shows effectiveness superiority over widely algorithms.