Clinical Decision Support Systems (CDSS) that utilization AI methods and their broadest feeling of man-made consciousness (AI) should be interpretable and straightforward. The absence of straightforwardness as opposed to offering help could rather turn into a variable of hesitation and obstruction. In this work, an extremely intricate and significant issue according to a clinical perspective is handled, to be specific the pathology known as Dry Eye Disease (DED), beginning from a case-control study on a HIV-positive populace and a solid piece of it. The contextual analysis is looked on two fronts, the first in which a gathering based bunching calculation is assembled. Besides, this calculation is separated to examine every part, making the investigation strategy straightforward and interpretable. In particular, a group of bunching calculations is introduced, like k-implies, agglomerative, unearthly, and birch, which are consolidated and utilized in two levels: in the primary, the marks are acquired from each clusterizer to perceive huge examples of the two populaces impacted by the DED pathology, within the sight of HIV and not. In this manner, the marks acquired at the main level are utilized as contributions on which the clusterizers are utilized once more, whose results in the last stage fill in as a preparation informational collection for an administered strategy (i.e., calculated relapse, choice trees, brain organization, and so on), to assess each and every part independently, using highlights significance methods (i.e., choice trees, LASSO relapse, Gini Importance (GI), Variable Importance (VI), and so on) Thusly, each bunching calculation utilized at the principal level can be viewed as another element in the following one and assess its singular commitment. Besides, every trademark is deciphered through explicit techniques for the significance of the qualities to settle on the choice help device as complete as could really be expected.
