Dataset features with ranges and measurement units.

We removed “diagnosis method” from the classification and feature selection phases, because it has the same values of “class of diagnosis” target we predict. We changed some feature names to add clarity: “blood lactic dehydrogenise (LDH)” into “lactate dehydrogenase test”, “cell count (WBC)” into “white blood cells (WBC)”, “cytology” into “cytology exam of pleural fluid”, “hemoglobin (HGB)” into “hemoglobin normality test”, “keep side” into “lung side”, “pleural glucose” into “pleural fluid glucose”, and “white blood” into “pleural fluid WBC count”.