Multiple Regression Based on Impact Parameters for Bruising Prediction in Apple

Abstract

Impact parameters were explored to be used for bruising predictions in apple by employing multiple regression analysis. All impact parameters observed were potential to be involved in the predictions except mass of fruit and impact duration. The multiple regression analysis based on maximum acceleration, velocity change, initial velocity, absorbed energy, maximum deformation and residual deformation produced a coefficient correlation (R) of 0.95 and 0.04% error for the relation between predicted bruise diameter and measured bruise diameter, a coefficient correlation (R) of 0.94 and 0.09% error for the relation between predicted bruise depth and measured bruise depth, a coefficient correlation (R) of 0.95 and 3.64% error for the relation between predicted bruise volume and measured bruise volume. The multiple regression based only on the maximum acceleration and velocity change still produced reliable bruising predictions.