A recursive approach for estimating missing observations in an univariate time series

Abstract

A linear recursive technique that does not use the Kalman filter approach is proposed to estimate missing observations in an univariate time series. It is assumed that the series follows an invertible ARIMA model. The procedure is based on the restricted forecasting approach, and the recursive linear estimators are optimal in terms of minimum mean-square error.

Keywords:

• ARIMA models
• Minimum-mean-square error
• Missing observations
• Recursive linear estimation