| Modern science relies more and more on dynamical properties. While the description of the state of a system is the first step to any scientific advance, the understanding of its dynamical components and how they evolve in time and interact with each other is usually a more complex, though more fruitful, step in an analysis. The analysis of such properties in the statistics framework is called time series analysis, the problem of grouping such processes in classes, and deciding in which group a dynamic process falls in, is called time series classification.; We here present applications of 2 time series expansion methods, Principal Component Analysis and Wavelet, first to the time series classification problem and then to the change detection problem (on-line time series classification).; We also present a new method to assert the behavior of statistical methods for variations around a hypothesized distribution using the Johnson system of distributions. |
