Unit 34 signal plus noise forecasts

34.1 Linear Signal

  • Fit \(b_0\) and \(b_1\) using least squares (or true values)

  • Find residuals

  • Fit an AR(p) to the residuals

Then find forecasts on the next noise value, generating an estimate of the signal (with the regression) and the noise (with the residuals). We will get more into this later and use some psis to find the limits.

34.2 Cyclic signal

Will do later, for now for both we can just do fore.sigplusnoise.wge For early lags, the noise behavior is dominant, and as we get further out, the signal dominates