An Alternative Approach to Estimate the Coefficients for Non-Orhogonal Data

Multicollinearity can cause serious problem in estimation and prediction, increasing the variance of least squares estimators of the regression coefficients and tending to produce least square estimates that are too large in absolute value. This paper shows how the dynamic programming algorithm can be used to estimate the regression coefficient. It also shows how the algorithm can be used to calculate various statistical quantities like the coefficient of determination, the F-statistics and the t-statistics.