USING FORECASTING ACCURACY CRITERIA TO DETERMINE OPTIMUM METHOD FOR ESTIMATING MISSING VALUES (AGRICULTURAL RESEARCHES DATA AS A CASE STUDY)

Abstract

Missing data in time series is considered as an important issue especially in the process of mathematical and statistical model estimation and consequently its forecasting. Arriving at certain results through the analysis based on the estimation of missing values by using different methods would have great effect on decisions based on the results especially in agricultural researchers. This requires studying the methods of missing data estimation represented by time series average method, arithmetic mean for adjacent values, median for adjacent values , linear interpolation and regression imputation and testing them by using forecasting accuracy criteria such as mean absolute percentage error (MAPE), mean absolute error (MAE), and mean square error (MSE) in addition to using simple regression models. Then comparing some statistical tests resulted from these results such as F- test , -test and - test to support the resulting results from forecasting accuracy tests to judge for the best methods in estimating missing values. The results of this research showed the suitability of regression imputation method according to its advantage in forecasting accuracy tests in addition to simple linear regression model tests which this research is advising to be used within the condition explained within the research text. It is necessary that data are suitable to get valuable results.