Ridge Estimation for Uncertain Moving Average Model Under Imprecise Observations

Abstract

By observing and studying the time series, looking for the law of its change and development, and predicting its future trend is time series analysis. The observation data in traditional time series are accurate, but the statistical data we get in actual observations such as carbon emissions and temperature are often inaccurate. We regard this type of imprecise observation as uncertain variables, model it with the uncertain moving average model (UMA), and estimate the parameters in the UMA model with the ridge method. Then through the residual analysis of the fitted model, we acquire the forecast value and the confidence interval. Finally, numerical examples are used to prove the feasibility of the method.