Robust Estimation for Semi-Functional Linear Model with Autoregressive Errors

It is well-known that the traditional functional regression model mainly based on least square or likelihood method. These methods usually rely some strong assumptions, such as error independence and normality, are not always satisfied. For example, response variable may contain outliers, term serially correlated. Violation of assumptions can result in unfavorable influences estimation. Therefore, a robust estimation procedure semi-functional linear with autoregressive developed to solve this problem. We compare efficiency our method through simulation study two real data analyses. The conclusion illustrates proposed outperforms method, providing random errors follow heavy-tail distribution.