Estimate of Stability of the Dynamic Clustering Algorithm for Temperature Signals

Abstract—A dynamic clustering algorithm for temperature signals is considered. It is used for solving the problem of distinguishing climatic regions characterized by a certain homogeneous climate type within the boundaries of the region on the Earth’s surface. A numerical experiment implemented with the aim of estimating its stability is described. The reference signal model used in the numerical experiment is presented as a sum of harmonic components. Additive components of the noise are formed in the frequency region. In the numerical experiment, the noise magnitude and sample size are measured. Estimates of root-mean-square deviations and stability of the dynamic clustering algorithm are presented.