The Human Development Index (HDI) is an important indicator in measuring the success of national development. Central Java with a high population can be considered as an obstacle and a driver of development. To find out the factors that affect HDI, it is necessary to make a model. One of the statistical methods that can be used is multiple linear regression analysis. However, in modeling multiple linear regression there are assumptions that must be met, namely linearity, normality, homoscedasticity, non-autocorrelation, and non-multicollinearity. If the non-multicollinearity assumption is not met, then another alternative is needed to estimate the regression parameters. Several methods that can be used are ridge regression and stepwise regression methods. The best model selection is done by looking at the smallest Mean Square Error (MSE) value. In this study, ridge and stepwise regression were applied to Central Java HDI data in 2021 and the factors that influence it, namely life expectancy at birth, expected years of schooling, average length of schooling, per capita expenditure, percentage of poor people, and unemployment open. Based on the Variance Inflation Factor (VIF) value of more than 10, it can be concluded that there is a multicollinearity violation. Modeling with stepwise regression produces the best model, with the smallest MSE value. The R square model value of 0,99 indicates that the model is included in the criteria for a strong model.
Copyrights © 2022