Comparison of Partial Least Squares Regression and Principal Component Regression for Overcoming Multicollinearity in Human Development Index Model

One of the assumptions in ordinary least squares (OLS) estimating regression parameter is lack multicollinearity. If multicollinearity exists, Partial Least Square (PLS) and Principal Component Regression (PCR) can be used as alternative approaches to solve problem. This research intends compare those methods modeling factors that influence Human Development Index (HDI) North Sumatra Province 2019 obtained from Central Bureau Statistics. The result indicates PLS outperforms PCR term coefficient determination squared error