Genetic programming as a method to develop powerful predictive models for clinical diagnosis

In the field of medicine it is of vital importance to accurately predict the presence of a disease (diagnostic prediction) or the future occurrence of a certain event (prognostic prediction). Genetic programming provides a method to develop such prediction models in an optimal way. In this paper we discuss as an example the diagnostic prediction of pulmonary embolism (PE), and compare the method of genetic programming with the logistic regression technique, which is well-known in the medical field. Categories and Subject Descriptors J.3 [Life and Medical Sciences]: Medical Information Systems General Terms Algorithms, Performance Keywords Genetic Programming, Diagnosis, Prognosis, Logistic Regression, Pulmonary Embolism

• Type: Conference
• Year: 2005
• Where: GECCO
• Authors: Ivar Siccama, Maarten Keijzer
• Keywords: Diagnostic Prediction, GECCO 2005, Genetic Programming, Pulmonary Embolism