Hi, Does anyone know how to include observation errors in the arima of R, which is implemented with the Kalman filter. I want to estimate observational error variance for noisy data in the context of ARMA model using arima of R. I read the manual and tried the example codes, but did not find the solution. From the outputs of the components "model", it seems to me that the default setting of the arima does not include the observational error in the fitting. The elements of matrix "H" are zeros when executing the example codes. Am I right on this one? Thanks in advance. Sincerely, Guiming Wang Natural Resource Ecology Lab Colorado State University Fort Collins, CO 80523 [[alternative HTML version deleted]]