The influence of segment size, on which the time series sample is divided, on the effectiveness of spectral analysis of the signals via eigenstructure methods with preprocessing of the sample by SSA method or modified SSA method is analyzed in the paper. The conditions of selection of segment (window) size for the modified SSA method when it is more effective as compared to initial SSA method is showed.
noise reduction in observation, eigenvectors, eigenvalues, singular vectors, singular values, modified SSA method, segment size, general statistical analysis
"Эffektyvnost spektralnoho analyza sobstvenno-struturnыmy metodamy pry predvarytelnoi obrabotke syhnalov modyfytsyrovannыm metodom SSA" [Effectiveness of spectral analysis by eigenstructure methods with signal preprocessing by modified SSA method],
Information Processing Systems,