Method and apparatus for classifying a time series of at least one second and the time series of the first

摘要

So as to examine the statistical dependence of their series (57) [Abstract] At conditional probability of the sequence is modeled when one by one neural network, respectively, to train a neural network. To determine the time-series proxy by using the neural network, to obtain the measure of the statistical dependence of the sample values ​​of the time series representative sample value of each series. Comparing with each other a measure of these dynamic series examined whether represented by the Markov process (n _N, n _1, ...) when the order of individual underlying.