ESTIMASI PARAMETER REGRESI RIDGE BINOMIAL NEGATIF UNTUK MENANGANI MULTIKOLINEARITAS; (ESTIMATION OF NEGATIVE BINOMIAL RIDGE REGRESSION PARAMETER TO HANDLING MULTICOLLINEARITY)
Ristyati, Awwalina Ghaida, Subanar
2015 | Skripsi | FMIPA UGMNegative binomial regression is a method used to analyze the relationship between response variable and explanatory variables, where response variable is count data. Commonly, maximum likelihood (ML) estimator is used to estimate parameters in negative binomial regression model. However, those estimation become insignificant if explanatory variables are collinear or multicollinearity occurred. Therefore, negative binomial ridge regression (NBRR) estimator is used to estimate parameters in negative binomial regression model in order to handling multicollinearity. In this thesis will be discussed about the modeling of number of unemployed in Semarang city and factors that influence it, such as inflation rate, GDP rate, minimum wage, and number of productive age population, based on negative binomial regression which involve multicollinearity problem. So that, to handling those multicollinarity will be used negative binomial ridge regression estimator. Mean squared error of maximum likelihood estimator and negative binomial ridge regression will be compared. The result indicated that the negative binomial ridge regression estimator is preferred to the maximum likelihood estimator.
Kata Kunci : negative binomial regression; maximum likelihood; ridge regression; mean squared error; multicollinearity
