Efficiency of Nonparametric Estimators for Missing Observations of Bilinear Time Series with Gaussian Innovation

Abstract

A time series is defined as data recorded sequentially over a specified period. Since the data are records taken overtime, missing observations in time series are very common. They may occur as a result of lost records, deletion of outliers, calender effects and defective measuring instruments Being unable to account for missing data has several limitations: A severe miss-representation of the phenomenon under study