<![CDATA[Maternal mortality is still a problem in all countries in the world, especially in developing countries like Indonesia. Several factors causing death are case detection, the risk of danger signs in pregnant women is still low, resulting in late case handling. Detection of danger signs of pregnancy is currently still done manually, but several studies have developed it machine learning Because it provides high accuracy, in this study researchers classified the detection of danger signs of pregnancy using an algorithm with techniques decision tree. Prediction of early detection of danger signs of pregnancy with 92% accuracy using comparative accuracy on 10 individual classifications (Nearest Neighbors, Decision Tree, Random Forest, Neural Net, AdaBoost, Gaussian Naïve Bayes, Bagging, Extra Tree, Gradient Boosting, Stacking) in this application a human in the loop was also developed for accuracy in providing recommendations. In predicting pregnancy danger signs, several studies use machine learning, because it has been proven to provide higher accuracy, interaction in providing predictions based on machine learning combined with expert intelligence, so that it can utilize big health data to solve problems and provide diagnoses and treatments. Test results, obtained values p<0.005 which mean Ordinal regression models can be used to predict maternal risk. Patients who are older, have more parity, lower height, distance between children < 2 years, HB < 11 gr/dl, LILA < 23.5 cm, have HBS-Ag, have HIV, have a history of DM, have a history of HT, positive for urine protein, hypertension and other diseases have a greater chance of having a high maternal risk. Conclusion, Application HITL can be developed for early detection of danger signs in pregnancy and provide appropriate recommendations to pregnant women and can determine high-risk pregnancies correctly, making it easier to care for and plan the place of delivery.]]>
Copyrights © 2024
