Machine Learning-Based Intrusion Detection System For Healthcare Data

Abstract

The rising advancement of intrusion strategies has given the desperate imperative for designing and developing IDS with excellent efficiency. The existing IDS have been developed to utilize obsolete threat datasets, concentrating too much on accuracy rate and less on prediction. Machine learning has the potential to deliver an efficient approach when it arrives at intrusion detection due to the high dimensionality and eminent dynamic nature of the available data in such mechanisms. However, plenty of the existing health care IDS either uses dynamic network performance measures or clients' biometric information to establish their database. This research introduced a NIDS for health care data using the Hybrid Feature Selection algorithm (Least Squares and Support Vector Machine), which minimizes forecast latency without influencing attack prediction efficiency by reducing the IDS complexity. The experimental results demonstrate the performance of the proposed hybrid method over the existing method in terms of precision, accuracy, recall, and F-measures.