Recursive Residuals Estimation for SEIV Models. Kalman Filter Method

      Аннотация: This book is concerned with a special method of estimation called recursive residuals that is applied for one of the forms of measurement error models, namely Structural Errors-in-Variables model. Kalman Filter technique is used to estimate the highlighted model, where Maximum Likelihood Estimation (under normality) is used to provide initial values. There are five proposed methods for initial value of latent variable which are recommended for specific conditions. Cumulative sum (CUSUM), that is a sensitive control chart tool, is involved to control the model fitting and to check the model assumption. Finally, real data applications are shown for environmental and economic domains.

Ключевые слова: Kalman Filter, Recursive Residuals, Cumulative Sum, Structural Errors-in-Variables model, Measurement errors model.