Abstract—Birth asphyxia is a major public health problem in the maternal and child health. It is the cause of illness, death or disability of a newborn baby. If doctors and staff have awareness to prevent and provide the proper treatments in timely manner, it will affect the quality of life of children in long-term. The purpose of this research is to predict birth asphyxia occurring using three base classifiers; Backpropagation Neural Network (BPNN), Support Vector Machines (SVMs), and Decision Tree (DT). Moreover, the popular ensemble learning, AdaBoost, also applied with the three base classifiers to improve their performances. The data used in this research were birth asphyxia data collected from Chaoprayayomraj Hospital, Thailand during 2006 – 2011. The results showed that DT model gives the best performance in all evaluation measures. However, AdaBoostBPNN model, instead, gives the best improvement with the accuracy of 87.80%. This model can be used to guide doctors and staff for preparing intensive care in special cases to prevent birth asphyxia occurring and reduce the rate of death and disability of the newborn.

Index Terms—birth asphyxia, classification, ensemble classifier, AdaBoost method

Cite: Punnee Sittidech, Nipaporn Chanamarn, and Kanokwan Arunrudchadarom, "Birth Asphyxia Classification Using AdaBoost Ensemble Method," Journal of Medical and Bioengineering, Vol. 4, No. 2, pp. 126-129, April 2015. Doi: 10.12720/jomb.4.2.126-129