Noncausal AR(2) AR(2) with one root in
and the other inside the unit circle
linear time series, uncorrelated but dependent This simulates a non-causal linear time series that
is uncorrelated but dependent (innovations need to be
non-Gaussian). Make an ACF of the time series and its absolute and
compare them.
ARCH(1) (refer to class notes for region of
stationarity)
GARCH(1,1) (refer to class notes for region of stationarity)
Bilinear(1,1) (refer class notes for region
of stationarity)
Markov Switching modelling
Switches between AR(1). Play with the parameters to see what
parameters allow for a stationary causal solution.
Integer Valued AR Order 1 and 2 (beware
of specifying parameters which stay within the stationarity zone for
order 2)
Poisson autoregression Again
care needs to be taken on ensuring the parameters lead to a
stationary casual solution).