Deteksi Dini Penyakit Diabetes Menggunakan Algoritma Neural Network Berbasiskan Algoritma Genetika

Primandani Arsi (STMIK Amikom Purwokerto)
Oman Somantri (Politeknik Harapan Bersama Tegal)

Jurnal Informatika: Jurnal Pengembangan IT
Vol 3, No 3 (2018): JPIT, September 2018

Diabetes or Diabetes Mellitus (DM) is a disease in which the glucose content in the blood can not be processed by the body. Based on the case of many diabetic patients, first action is needed as a solution of the problem of diabetes by predicting to detect early diabetes. This is necessary because often medical decisions are made based on experience and rational reasoning. Prediction of diabetes can be done by using some data of diabetic patients who have been stored in the database to make a pattern for the determination of diabetes with Artificial Intelligen technique so that the result of inaccurate diagnosis can be avoided. In this study the authors apply Genetic Algorithm (GA) to optimize Neural Network (NN) model by searching the best parameter value on Neural Network (NN) model. The experimental results showed a decrease in RMSE value which means an increase in predicted accuracy value, ie from 0.402 +/- 0,035 to 0.396 +/- 0,032.The optimization of the NN model prediction, the relevant policy makers can determine the prediction of diabetes more accurately.