The prediction of heavy-tailed autoregressive sequences of first order is considered. Necessary and sufficient conditions are obtained for the linearity of the regression predictor with time reversed, which answers a question raised in [13]. Also, the error of the linear regression predictor is compared with that of the best linear predictor in the symmetric stable case.
No takes yet. Share an insight, caveat, or question.
Cambanis et al. (1995) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: