<![CDATA[Diabetes mellitus is one of the leading causes of death a disease because it can cause many health complications. Method that can help to diagnose this disease early is sorely needed. Genetic programming is a method used in this research. Genetic programming is one evolutionary algorithm that uses parse tree as its solution representation. This method will produce decision tree as its output which will be used to diagnose patients in the testing dataset. This research will also observe the correlation between genetic programming parameters and fitness value. Tree with highest fitness value produced with population number 900, maximum iteration 300, crossover rate 0.8, mutation rate 0.1 and training data with ratio of patient with diabetes and patient without diabetes 1:2. Decision tree produced with those parameters will be used to diagnose diabetes patient in testing data. The accuracy of decision tree produced with this method is 66,11%.]]>
Copyrights © 2019
