A simulation study of first-order autoregressive to evaluate the performance of measurement error based symmetry triangular fuzzy number

Abstract

Data collected by various data collection methods are often exposed to uncertainties that may affect the information presented by quantitative results. This also causes the forecasted model developed to be less precise because of the uncertainty contained in the input data used. Hence, preparing the data by means of handling inherent uncertainties is necessary to avoid the developed forecasting model to be less accurate. Traditional autoregressive (AR) model uses precise values and deals with the uncertainty normally in forecasting model. Fewer researches are focused on data preparation in time-series autoregressive for handling the uncertainties in data. Hence, this paper proposes a procedure to perform data preparation to handle uncertainty. The fuzzy data preparation involves the construction of fuzzy symmetric triangle numbers using percentage error and standard deviation method. The proposed approach is evaluated by using the simulation method for first-order autoregressive, AR (1) model in terms of forecasting accuracy performance. Simulation result demonstrates that the proposed approach obtains smaller error in forecasting and hence achieving better forecasting accuracy and dealing with uncertainty in the analysis.