Parallel Computing of Kernel Density Estimates with MPI

Abstract. Kernel density estimation is nowadays a very popular tool for nonparametric probabilistic density estimation. One of its most impor-tant disadvantages is computational complexity of calculations needed, especially for data-based bandwidth selection and adaptation of band-width coefficient. The article presents parallel methods which can signif-icantly improve calculation time. Results of using reference implemen-tation based on Message Passing Interface standard in multicomputer environment are included as well as a discussion on effectiveness of par-allelization. Key words: kernel density estimation, plug-in method, least squares cross-validation, adaptive bandwidth, parallel algorithms, MPI