In the diagnosis of disease, traditional oriental medicine stresses questioning of the patient, and adds further observations to reach a comprehensive conclusion. Patient questioning yields multidimensional data covering some 200 or more items, and while an ideal healthy subject, from the viewpoint of traditional oriental medicine, preferably gives all-zero data, a disadvantage is that in the inverse matrix analysis, the mean and standard deviation cannot be obtained in the unit space. As a solution to these problems, this study applied the MTA methodology, using an adjoint matrix. An analysis was made of the diagnosis of patient recognition of what are referred to as blood diseases in traditional oriental medicine, by defining the values of all items as zero for healthy subjects and using data with many zero standard deviations as a unit space. Accurate diagnostic results were obtained.
