Early detection of glaucoma can be achieved through prophylactic screening of population. This leads to a sharp increase in the patients' flow and deprives the physician of a possibility to examine them by routine methods. Such examinations will be feasible if to reduce the procedure time by automation, or to cut the examination program after the analysis of the existing medical programs by mathematical methods. The authors have used the method of linear discriminant functions which permits one on the basis of the medical data statistical analysis to learn the informativeness of diagnostic symptoms and decisive rules to form risk-groups of the patients who need more thorough medical examination. The analysis accomplished has indicated that there is no closely interlaced symptoms. Among most informative there are signs considerably differing in their statistical characteristics. The diagnostic rules proposed in this article provide adequately high accuracy in selection of healthy people groups and, as well, the patients with suspected glaucoma or suffering this disease.