Representation of Time Series
Basically, one can describe time series either in the time domain or in the frequency domain. Difference equations are used in the former, and frequency spectra or transfer functions are used in the latter to specify the dynamic structure of time series. Both representations are used in this book. This chapter discusses time series models first in the time-domain, then using transfer functions in the frequency domain. The time domain representation of time series can further be divided into two broad and different modes: the traditional and the newer Markovian, i.e., state space representation mentioned in Chapter 2.
KeywordsTime Series Kalman Filter State Space Model ARMA Model Seasonal Component
Unable to display preview. Download preview PDF.