The analysis and classification of birth data

Abstract

The paper presents a study regarding the births that took place at the Bega Obstetrics and Gynecology Clinique, Timişoara, Romania in 2010. The analysis began from a dataset including 2325 births. The article presents a synthesis of the studies that analyze birth data. The Apgar score is the main subject in many studies. On one hand, researchers investigated the relation between the Apgar score and different factors such as the newborns’ cry, the level of glucose in the blood from the umbilical cord, the mother’s body mass index before the pregnancy, etc. On the other hand, there are studies that demonstrate that the Apgar score is important for the ulterior evolution of babies. The article presents the attributes from the dataset and how they were preprocessed in order to be analyzed with Weka. The values of each attribute were investigated and the results were presented. The past experience regarding births, expressed through the dataset values, was then used to build classification models. With the help of these models, the Apgar score can be estimated based on the known information regarding the mother, the baby and possible medical interventions. The purpose of these estimations is consultative, to help in identifying which values of the input variables will lead to an optimal Apgar score, in certain circumstances. The classification models were built and tested with the help of ten classification algorithms. After the model that produces the best results of classification was determined, a dedicated application was developed, with the aid of the Weka API which classifies the birth data by using LogitBoost algorithm.