Figure 6.

Pattern recognition and probabilistic diagnostic plots. This figure was created by inputting the probability of classification of each individual in different groups, i.e., the individual's chances of belonging to different groups, given by the PLS models. In the star plot (left), each star is an individual, and the colors are their true diagnosis. The x and y axis represent the diagnosis according to our classification models. The point where the stars cross the circles in each axis represents the probability of an individual's being labeled as having the diagnosis coded by that axis. The fact that the stars are elongated where the color (true diagnosis) agrees with the axis diagnosis indicates that the vast majority of patients are correctly classified. At right, a different representation of the same data, easier to visualize the probability of diagnoses in a single individual. Now, the colors represent the diagnosis given by our classification models. Each line represents an individual; the crossing point between the colored lines and the axial lines represents the probability of such an individual's being given that diagnosis. The four small panels at right show the probability curves of diagnosis for four selected individuals (bold arrows), color-coded by the true diagnosis. Y axis ranges from 0 to 1 and encodes the chance of the selected individual of being classified, by the algorithm, with the diagnosis in the X axis. For instance, the individual in the upper left quadrant has almost no chance (close to 0) of being classified as AT, a low chance of being classified as HD, a higher chance of being classified as control, and a high chance of being classified as AD. In fact, this individual had AD, as revealed by the color (purple) that represents the true diagnosis.