Predicting The Risk of Preeclampsia with History of Hypertension Using Logistic Regression and Naive Bayes

The causes of maternal and infant mortality are high, especially in developing countries other than bleeding, namely preeclampsia because mild preeclampsia increases more sharply than severe preeclampsia. The incidence of preeclampsia varies in each country even in every region. The purpose of this study was to compare two methods in data mining techniques, namely the Logistic Regression method and the Naive Bayes method, to predict the risk level of Preeclampsia on the seventeen existing attributes data. The data obtained in this study were 239 preeclampsia patients who were obtained from Surabaya General Hajj Hospital, by collecting data from 2016 to 2017. This study used performance measurement of the two methods using score with the ROC curve (Receiver Operating Characteristics). In the experimental results, we found that in the term of accuracy and ROC curve, Logistic Regression has a slightly better performance than the Naive Bayes method.