Techniques of Time Series Modeling in Complex Systems.

The autoregressive moving average (ARMA) and the autoregressive integrated moving average (ARIMA) are important techniques for time point analysis. An ARMA is a stationary model and by taking a series of differences, it is necessary to apply it to stationarity. Data can be classified as serially correlated using three basic criteria: mean, variance and covariance. Detrending is a strategy of removing the trending components from the time series. Detrending data may be used to see subtrends when the data shows an overall increase. Differencing, SARIMA, Croston’s intermittent demand forecasting approach and bagging are other statistical methods developed by unstable decision rules to improve the accuracy of forecasts. Time point forecasting is abbreviated as ETS models or exponential smoothing state-space models considering error, trend and seasonality. Here we try to address time point analysis techniques and parameters in complex systems.