Predicting Fetal Condition from Cardiotocography Results Using the Random Forest Method Syifa Fauziyah Nurul Islam (a*), Intan Nurma Yulita (b)
a,b) Department of Computer Science, Padjadjaran University, Sumedang 45363, Indonesia
*syifafauziyah899[at]gmail.com
Abstract
Cardiotocography is an important process in pregnancy as fetal monitoring. It monitors the babys heart rate in a healthy condition or not. Apart from that, this can also measure whether the movements carried out by the baby in the womb are normal or not. This study extracted the recording data by cardiotocographs. The attributes of fetal data that have been recorded amount to 22. They were used as the indicators in determining the conditions of the fetus whether under normal circumstances, suspect or pathologic. The prediction of the fetus condition was based on the Random Forest method. Also, the method was compared with the Naïve Bayes and Decision Tree methods. The accuracy of the Random Forest method reached 95.12%. It was higher compared to using other methods.
Keywords: Cardiotocography; Radom Forest; Naive bayes; Decision Tree
