Dear All R users,
I have many time series. Each of them follows the
structure below,
x_t = a_0 + a_1 x_{t-1} + a_2 x_{t-2} + \epsilon_t if x_{t-d} < \theta
b_0 + b_1 x_{t-1} + b_2 x_{t-2} + \epsilon_t if x_{t-d} >= \theta
I hope to classify these series into a few groups. Assume
the delay parameter d is same among all the series but
the threshold parameters \theta are different. Also the
parameters a_i,b_i of the series are same if these series
belong to the same group but are different if the series
belong to different group. Can you give me some advices in
classifying these series by cluster analysis. Thank you in
advance.
--
Jin