Figure 3.

Classification of those who belonged to the clinical group in model

Note. The validity of Discriminant Analysis is tested by contrasting the power to discriminate provided by the Discriminant Model with a chance classification: reclassification of the Female Cancer Group (n= 236) and of the Control Group (n= 232) obtained with the Discriminant Model. On the left we have reported the “improvement over chance criterion” (IOCC), showing that more that 80% of the Female Cancer patients are correctly reclassified by the Discriminant Function (gray bar, left). On the right re-classification is obtained with the cross-validation procedure. It is described as a ‘jack-knife’ classification, in that it successively classifies all cases but one to develop a discriminant function and then categorizes the case that was left out. This process is repeated with each case left out in turn. This cross validation produces a more reliable function. Here the difference between the results obtained with the two methods is minimal, as more than 79% of the Female Cancer patients are correctly reclassified with cross-validation (gray bar, left). Both procedures detect a valid discriminative power for the Model