A Machine Learning Approach to Predict NICU Units in Hospital by Developing Smart e-Healthcare System

With the advent of latest technology, the health care system of our society needs smart medical framework. They are using smart technologies to combat the orthodox offline health care system by providing prompt medical services. This research work reviews the several related literatures and gives an insight to the requirement of Neonatal Intensive Care Unit (NICU) beds in hospitals for preterm babies in emergency sector. A smart e-health care framework is developed that estimates the availability of NICU units in and around the vicinity of the person to avail the proposed service. The framework consists of an application developed using android that the users can use to interact with the hospitals in crisis hours. Cloud storage has been used to store the data securely using the AES algorithm. Cloud storage also eases the accessibility of the data and is reliable. The proposed framework uses machine learning algorithms to predict the availability of NICU units in hospitals for newborn preterm babies efficiently. The CNN model extracts the essential features from the dataset, and SVM performs the classification task. Moreover, the number of available units in a hospital has also been reflected in the application developed in real-time basis, so that the users can accordingly contact the respective hospital in emergency situation. The proposed framework has outperformed the earlier CNN and SVM models with 90.4
^%
recall, 90.6% precision, and 95.4% accuracy.