Fig. 5. Probability densities of PS^i,tfor the observation vectors of the initial complex. Vectors are defined with table 7, PS^A= (2,1,2,0,1), PS^B= (0,1,1,3,1), etc.  The first place in the observation vector PSⁱi represents the number of patients who rated the combination i with the first score, the second place represents the number of patients who rated the combination i with the second score, and so on. The cutting point for the partitioning of the complexes occurs at the point of the highest probability, between the combination F and G with P (PS^F< PS^G= 0.73).