Jurusan Teknik Informatika, STT Nurul Jadid Probolinggo
Email : email@example.com
Diabetes is a group of chronic diseases that can cause death . Communities particularly hard layman to detect because the symptoms of diabetes are not clearly visible coupled with long duration and lack of accurate detection equipment only detects diabetes if blood glucose levels . In the detection of diabetes research has been done that ANN with back propagation algorithm method and Fuzzy logic based intelligent system . However, this method is still considered optimal because the average results that show its accuracy ata rate less than the maximum . The new method is proposed with a view to finding a solution better than previous methods in terms of average accuracy , which is the optimization of the neural network algorithm . Neural network is used as a method of artificial intelligence to detect diabetes , while the genetic algorithm is used for optimization of neural network . Evaluation and testing methods used are cross validation , and confusion matrix . Cross validation test results show that the neural network optimization method using a genetic algorithm produces high accuracy is 98.57 % better than without optimization neural network method which produces 94.29 % accuracy and also better than Fuzzy logic based intelligent system that produces accuracy 97 %.
Keywords: Detection of diabetes, neural network and genetic algorithm