Improving Prediction of Peroxide Value of Edible Oils Using Regularized Regression Models