Adaboost Based Risk Level Prediction Model For Maternal Health

Most of the pregnant women are ignorant about the risks factors and die from complications of pregnancy. Therefore, the health data of pregnant women should be constantly monitored. Risk levels should be determined. In this study, machine learning techniques were used to automatically determine the risk levels of pregnant women in order to reduce the workload of doctors. Health data of pregnant women were analyzed to determine the level of risk intensity. Classification approaches according to the intensity of risk were used for the analysis of risk factors. Decision tree and Adaboost algorithms were used in the study. Among the decision tree algorithms, Random Forest, Rep Tree, Random Tree and SimleCart algorithms were preferred. The data were first classified using only decision trees. Then, the Adaboost algorithm applied data was classified using decision tree algorithms. The results were compared. As a result, it was seen that Adaboost algorithm increased the classification success. The highest accuracy value was obtained when Adaboost and Random tree algorithm were used together. This value is 86.09%. For this reason, a model in which Adaboost and Random tree algorithms are used together is proposed.