Hello, I have a couple of Qs about arima() in ts package. 1. I know the basic concept of arima model, but not very sure about the value returned by arima(), such as sigma2. from the help in R, sigma2 is the MLE of the innovations variance.. but i am still not sure what this "innovations variance" refer to. My Q is which book or paper I should look at to figure out the terms used by arima() function? The help in R on arima listed quite a few reference. I am wondering which one I should look at to get a quick hold of the meanings of the input and output parameters of arima()? 2. How to determine the input parameters of arima(), like *order*? Is it a trial-and-error process? I try a few of them, then pick a best fit one based on the model fitting results, i.e., the returned value. 3. My objective is prediction(fine grain interpolation), e.g., if I interpolate a new point for every two original points in the original series, after model fitting, I have the coefficient of different components in the model. Is there a function that can output the interpolated values based on returned value of arima()? THANKS a lot in advance! yan