<![CDATA[Poverty is a multidimensional phenomenon influenced by various socioeconomic factors that are often correlated with one another. These strong relationships among variables can trigger multicollinearity problems in conventional regression analysis. This study aims to compare the effectiveness of the Ridge Regression and Partial Least Square (PLS) methods in overcoming multicollinearity issues in modeling the factors influencing poverty in Simalungun Regency from 2017 to 2025. This study utilizes secondary time series data, with the Poverty Rate (Y) as the dependent variable and independent variables including the Human Development Index (X_1), Open Unemployment Rate (X_2), Gross Regional Domestic Product (X_3), Population Density (X_4), and Total population (X_5). Initial detection results indicate the presence of strong multicollinearity, where the population density and Total population variables have Variance Inflation Factor (VIF) values greater than 10. Based on the evaluation of the model's goodness of fit, Ridge Regression proved to be more optimal than PLS in addressing this multicollinearity problem. The Ridge Regression model yielded a high Adjusted R-Square value of 88.93% and a Mean Square Error (MSE) of 0.0369. On the other hand, PLS only produced an Adjusted R-Square of 57.20% and an MSE of 0.1426. The partial testing results on the best model (Ridge Regression) concluded that the Human Development Index (HDI) and the Total population have a significant negative effect on the poverty rate in Simalungun Regency.]]>
Copyrights © 2026
