PURPOSE:To analyze time-series data on the basis of nonlinear dynamic theory by taking versatile analyses such as a track display analysis, a dimension analysis, a Liapunov spectrum analysis, an entropy analysis, and a decision theoretic nonlinear predictive analysis, and totally processing the data. CONSTITUTION:After the time-series data are inputted 1, embedded parameters are set 2 and the inputted time-series data are converted 3 into track data in a reconstitution state space using time shift coordinates by an embedding process. Versatile analyses of the track data such as the track display analysis 4, dimension analysis 5, Liapunov spectrum analysis 6, entropy analysis 7, and decision theoretic nonlinear predictive analysis 8 are taken. Then, total processing 9 is carried out by mutually utilizing the respective analytic results and the processing after the embedded parameter setting processing 2 is repeated when necessary. When it is judged that the total analysis 9 is completed finally, the total analytic result is outputted 10.
展开▼
