PERBANDINGAN METODE KLASIFIKASI REGRESI LOGISTIK BINER DAN RADIAL BASIS FUNCTION NETWORK PADA BERAT BAYI LAHIR RENDAH (Studi Kasus: Puskesmas Pamenang Kota Jambi)

Published: .
Open Access

Citation Format:
Abstract

Low Birth Weight (LBW) is one of the main causes of infant mortality. LBW must be identified and predicted before the baby birth by observing historical data of expectant. This research aims to analyze the classification of status newborn in order to reduce the risk of LBW. The statistical method used are the Binary Logistic Regression and Radial Basis Function Network. The data used in this final project is birth weight at Pamenang Jambi City health center in 2014. In this research, the data are divided into training data and testing data. Training data will be used to generate the model and pattern formation, while testing the data is used to measure how the accuracy of the representative model or pattern formed in classifying data through confusion tables. The results of analysis showed that the Binary Logistic Regression method gives 81,7% of classification accuracy for training data and 77,4% of classification accuracy for testing data, while Radial Basis Function Network method gives 92,96% of classification accuracy for training data and 80,64% of classification accuracy for testing data. Radial Basis Function Network method has better classification accuracy than the Binary Logistic Regression method.

 

Keywords: Low Birth Weight (LBW), Binary Logistic Regression, Radial Basis Function Network, Classification, Confusion

Keywords: Low Birth Weight (LBW); Binary Logistic Regression; Radial Basis Function Network; Classification; Confusion

Article Metrics: