Rough set is a mathematical approach to study imprecise, uncertain or incomplete information or knowledge. The data mining technology based on the rough set theory can be trained through the training data set. The technology is applied to such fields as the prognosis of the pain in the chest, emergence diagnosis in ICU, distinguishing and diagnosis of lung tumor and rheumatoid disease, diagnosis of galactophore disease, diagnosis of premature birth, classification and diagnosis of liver disease and acute appendicitis. The technology has a brilliant future in medical data mining.

Based on the statistical materials of SARS epidemic process in a general hospital in 2003 and with least square method, the epidemic laws are fitted through MATLABL, and the infection characteristics and relative conclusions of SARS are analyzed.