KLASIFIKASI DATA BERAT BAYI LAHIR MENGGUNAKAN PROBABILISTIC NEURAL NETWORK DAN REGRESI LOGISTIK (Studi Kasus di Rumah Sakit Islam Sultan Agung Semarang Tahun 2014)

Abstract

Birth Weight Infant (BWI) is the baby’s weight weighed in an hour after being born. Factors that may influence the BWI such as maternal age, length of gestation, body weight, height, blood pressure, hemoglobin and parity. One possibility of BWI is Low Birth Weight Infant (LBWI) (BWI < 2500 gram). LBWI is one of the causes of infant mortality. This study use the Probabilistic Neural Network (PNN) and Logistic Regression to classify the birth weight of infant in RSI Sultan Agung Semarang along the year of 2014. This study’s aims are to know the factors that affect the BWI by using logistic regression and finally finding the best method between PNN and logistic regression methods in classifying the BWI data. As a result, gestation, body weight and hemoglobin are the factors that affect the BWI in RSI Sultan Agung Semarang. The accuracy of PNN classification method on training data is 100%, which is better than the logistic regression method giving only about 88,2%, while the testing data has the same great accuracy at 86,67%.

Keywords: BWI, LBWI, PNN, Logistic Regression, Classification