Using Robust Ridge Regression Diagnostic Method to Handle Multicollinearity Caused High Leverage Points

Statistics practitioners have been depending on the ordinary least squares (OLS) method in linear regression model for generation because of its optimal properties and simplicity calculation. However, OLS estimators can be strongly affected by existence multicollinearity which is a near dependency between two or more independent variables model. Even though presence estimate still remained unbiased, they will inaccurate prediction about dependent variable with inflated standard errors estimated parameter coefficient It now evident that high leverage points are outliers x-direction prime factor collinearity influential observations. In this paper, we proposed some alternative to methods estimating multiple cause problem. This procedure utilized estimates as initial followed an ridge regression. We incorporated Least Trimmed Squares (LTS) robust down weight effects lead reduction multicollinearity. The result seemed suggest RLTS give substantial improvement over Ridge Regression.