Knowledge discovery in medical databases: what factors influence a successful bone marrow transplant for Hodgkin's disease

The paper has given a general introduction into the methods which can be used in determining which factors influence a successful bone marrow transplant for the treatment of patients with relapsed or resistant Hodgkin's Disease. By following the KDD process and applying various data reduction, data cleansing and data mining techniques to the original EBMT database several nuggets of knowledge have been created. Firstly the induction rules may provide the basis for predicting the outcome in patients depending upon their characteristics. The artificial neural networks shows which characteristics of the patients are the most relevant, and the association rules present the characteristics which generally appear together for the different outcomes.