METHOD AND DEVICE FOR PREDICTION OF NON-STATIONARY TIME SERIES

摘要

1. A method of predicting nonstationary time series based on the autoregressive moving average model with a minimum mean square error, characterized in that the recycling step further comprises input data by one sample, an oversampling of the original process a logarithmic scale of time, finding the energy spectrum of the received signal, determining the response, corresponding entropy energy spectrum of the sample, calculate the maximum value of the entropy responses residence the forecast for the implementation corresponding to the maximum entropy oversampling results in exponential forecasting time. ! 2. Device prediction nonstationary time series, comprising analog-to-digital converter, a synchronization unit coupled to the second input of the analog-digital converter, a first input of which is the input device, the predicting unit, characterized in that it further comprises a delay unit, whose first input is connected to synchronization output unit, a second input connected to the output of analog-to-digital converter, an output coupled to the input unit of logarithmic resampling process starting time scale, the output of which is connected to the first input buffer memory unit connected to its first output to the input of block finding signal spectrum, the output of the latter is an input unit for calculating the square of the signal module, the output of which is connected to the input of finding sample entropy block, whose output is connected to the input of block calculating maximum values