Application of Uncertain Autoregressive Model to Per Capita Output of Yarn in China

Abstract

This paper aims to employ time series analysis to forecast the per capita output of yarn in China. We note from random time series analysis that the residual data did not pass the normality and homogeneous distribution tests. It means that the disturbance term cannot be assumed as a random variable. The random autoregressive model is not suitable. In this situation, this paper employs an uncertain autoregressive model in which the disturbance term is an uncertain variable. Besides, an uncertain hypothesis test is designed to evaluate whether the uncertain autoregressive model fits the observed values well. From this study, we conclude that the uncertain autoregressive model is a good fit for the per capita output of yarn.