In a few cases especially in economic and social sector there are a lot of data structure which the response variable has limited values or missing data that was called censored data. Into censored data the best regression model estimation using OLS (Ordinary Least Square) would obtain biased parameter estimation result, because there are zero values that makes the assumption of total error will not be zero.
Tobit regression analysis method can be used to solve lost information in Logistic Regression and Probit Regression such as the data will lose information like the variance of response variable. The algorithm used in computation arithmetic of Tobit Regression is EM-Algorithm (Expectation step and Maximization step).
Based of the result of analysis, It can be concluded that the relationship between response variable Y (amount of sugar production SHS) with explanatory variables X (X1: weights of sugar cane mill, X2: Pol molasses, X3: Pol Yield, and X4: molasses weights include raw sugar) is :
= 366.4458 + 11.9257 X1 – 33.0503 X2 + 101.7885 X3 + 36.9823 X4
with highest pseudo-R2 0.2755
Filed under: Statistika Ditandai: | TOBIT REGRESSION ANALYSIS FOR THE SHS (SUPER HIGH SUGAR