Modeled early detection of pregnancy risk based on Poedji Rochjati score card using relief and neural network
The safety and healthy of pregnant women and their babies is very important. An anticipatory action should be prepared as early as possible to prevent or reduce the high risk of pregnancy. The Poedji Rochjati Score Card (PRSC) is one of the methods that can be used to know the pregnancy risk used by doctors and midwives. This research proposed modeling PRSC using artificial neural network (ANN) method and select the most important factor in determining the pregnancy risks using ReliefF algorithm. The results of early pregnancy risk detection using ANN is expected to assist the process of checking the risk of pregnancy, either by pregnant women or by health workers. Experiment show that the best configuration was using 4 neighborhood parameter of ReliefF algorithm and 5 hidden neuron parameter of Neural Network. And the significant feature are bleeding during pregnancy, baby dies in uterus, never failed pregnancy, had caesarean section, too late pregnant, first pregnancy when age 35th, age<= 16th, too soon pregnant again, diabetes, blood deficiency, had given birth and was given infusion / transfusion, age> = 35th, pregnant twins 2 or more.