DeepNS: A Deep Learning Approach to Predicting Early-Onset Neonatal Sepsis

Research Supervisor
Journal Title
Journal Title
Journal ISSN
Volume Title
Neonatal sepsis is a major concern for maternal and neonatal health which remains a global problem with little progress made despite major efforts. The subject of this work is to design a non-invasive, deep learning classification model for predicting early-onset sepsis in neonates in Neonatal Intensive Care Units accurately and efficiently. By non-invasive, it means that no external instrument or foreign body is introduced when taking data for the classifier. Moreover, the data collected for the purpose of predicting and classifying subjects with neonatal sepsis is in the form of tabular, structured data (comma separated values, CSV). The deep learning classification models we design and propose in this work (such as Artificial Neural Network, Convolution Neural Network and Recurrent Neural Network) are known for working with time series, sequential or image data. Hence, the aim of this work is to develop such a model that uses the powerful tools inherent in Neural Networks for pattern recognition, and use them to outperform traditional machine learning algorithms in order to predict early-onset neonatal sepsis. Additionally, we design other machine learning classifiers as well to have a fair comparison between machine learning and deep learning model to show Neural Networks are able to trump the performance of these classifiers. Real life neonatal sepsis data samples from two different hospitals are used (Crecer’s Hospital Centre in Cartagena-Colombia and Children’s Hospital of Philadelphia) to make the evaluation of the Neural Networks as authentic as possible.
Department Name
Electrical and Computer Engineering
North South University
Printed Thesis