Enhanced Security for a Patient Healthcare Delivery Realtime Vital Signs Monitoring and Alert Ensemble

Abstract

Remote health monitor systems have enormous potential of becoming an integral
part of medical system. Their outstanding roles are seen in treatment and monitor
of patients with critical healthcare issues to reduce unnecessary visits to hospitals
and unneeded pressure of healthcare experts. Health care monitors generate
enormous data that must be analyzed to aid improved care delivery. We thus,
advance a deep learning deep-learning techniques to detect reliability and
accuracy of data obtained via an IoT-based remote health monitor. With dataset
retrieved from Kaggle, we seek minimum training error that will also result in the
best fit, selecting the number of hidden layers (a neuron for each layer) was
established via a trail-and-error method, and examining the results. The best
possible number of layers was found via tests on single layer with 1-to-20 neurons,
and shows that our best F1-score with the least amount of train-loss time is with
the configuration of 9-neurons and F1 of 93% and train time loss of 1.140. The
accuracy comparison is performed between strongly correlated features and
weakly correlated features. Finally, accuracy comparison between two approaches
is performed to check which method is performing better for detecting erroneous
data for the given dataset.