The Parameter Estimation of Time-varying GM(1,1)

Abstract

In this paper, we proposes a Time-Varying GM (1,1) (TVGM(1,1)), and using recursive least square with forgetting factor to estimate model parameters. The TVGM (1,1) provide adaptation parameter for the data does not unity environment. Which we can track system response well. From computer simulation indeed show the excellent performance, and what's more, the smoothing method and background value method for preprocessing data in the traditional GM(1,1) can be also applied to the proposed TVGM (1,1) to increase the system reliability and extend the application fields.