METHOD AND DEVICE FOR ASSESSING MULTIVARIABLE TIME SERIES DATA ON LIVING BODY OR THE LIKE

PROBLEM TO BE SOLVED: To provide a general-purpose method for grasping correlations between plural time series phenomena as system elements, for quantitatively and objectively assessing system functions related with a living body or living body activities, and for quantitatively understanding the degree of freedom in the system, the cooperativeness and stability among the elements in the system, based on the correlations obtained. SOLUTION: Using plural periodic time series data on living body signals, correlation coefficients between the time series data obtained on a periodical unit basis are calculated as correlation matrices. For every element in those correlation matrices, the mean value and the standard deviation among the plural periodic periods and the eigenvalue of the mean-value matrix of those correlation matrices are calculated. This mean-value matrix of each element in the correlation matrices represents the cooperation pattern, and this standard deviation matrix of each element in the correlation matrices shows the cooperation stability, while the number of eigenvalues exceeding one in value indicates the degree of freedom. These enable us to make a judgment on the cooperation characteristics about functions of the living body system or the like and their conditions in time-based changes. COPYRIGHT: (C)2006,JPO&NCIPI