INFLUENCE OF UNOBSERVED DATA IN THE TIME SERIES OF THE DEPENDENT VARIABLE AND THEIR POSITION IN ANALYSIS OF MULTIPLE LINEAR REGRESSION ON PREDICTION - CASE STUDY ON: FACTORS AFFECTING CO_2 EMISSIONS

Using a variety of statistical techniques, time series forecasting is crucial for preparing for and predicting the future. It is contingent on making an accurate forecast as to the value of a variable at some unknown time in the future. This research analyzed the missing data from the time series (a model with no missing observations and three models were considered to be missing data for the dependent variable at various positions). By a standard multiple linear regression of the four models of the study, it is clear that the series is consistent, transparent, within the bounds of statistical acceptability, the analysis used the Ordinary least square and the weighted least square to find the best prediction model with missed observation.