مقارنة بين طريقتي Yates و Rubin في تقدير القيم المفقودة في تصميم المربع اللاتيني

الملخص EN

In this research tackling the missing values in the Latin Square design and two methods (Rubin, Yates) and to compare these methods were used the Effect coding Regression and analysis of varianceand preference between these methodsused several statistical measures which the value of( the relative efficiency R.E, and coefficient of variation C.V and the mean square error MSe) data are drawn from the Faculty of Agriculture and Forestry , University of Salahhdin/ Erbil Agricultural research Department in the site CardaRasha.This experience compared to five rows and five columns and five varieties of wheat and every class represents the results are shown theRegression analysis method with using effect coding and analysis of variance method to correspond when not find missing value, and the relative efficiency and coefficient of variation equal tooin the case of the missing one value, the relative efficiency and coefficient of variationin the case of two missingvalue the same column in the Yates method are less than Rubin method ,and anther row ,treatment value .

thevalue of (MSe)in effect coding and analysis of variance method of areless all the case, the relative efficiency , the case of three missing values equally to efficiency and coefficient of variation of these methods.