It is well-known that the traditional functional regression model is mainly based on the least square or likelihood method. These methods usually rely on some strong assumptions, such as error independence and normality, that are not always satisfied. For example, the response variable may contain outliers, and the error term is serially correlated. Violation of assumptions can result in unfavorable influences on model estimation. Therefore, a robust estimation procedure of a semi-functional linear model with autoregressive error is developed to solve this problem. We compare the efficiency of our procedure to the least square method through a simulation study and two real data analyses. The conclusion illustrates that the proposed method outperforms the least square method, providing random errors follow the heavy-tail distribution.
18 February 2023 in Journal of Systems Science and Complexity
In many statistical applications, data are collected sequentially over time and exhibit autocorrelation characteristics. Ignoring this autocorrelation may lead to a decrease in the model's prediction accuracy. To this end, assuming that the error process is an autoregressive process, this paper introduced a semi-functional linear model with autoregressive errors. Based on the functional principal component analysis and the spline method, we obtained the estimators of the slope function, nonparametric function, and autoregressive coefficients. Under some regular conditions, we found the convergence rate of the proposed estimators. A simulation study was conducted to investigate the finite sample performance of the proposed estimators. Finally, we applied our model to forecast the monthly retail sales of electricity, which illustrates the validity of our model from a predictive perspective.
16 October 2014 in Journal of Systems Science and Complexity