Intrusion detection system for Healthcare based on Convolutional Neural Networks

The rapid growth and development of the Internet of Medical Things (IOMT) has significantly altered how diseases are managed, improved ways for diagnosing and treating diseases, and decreased the cost and errors associated with healthcare and as a result, network security has always been a crucial problem especially when it comes to maintaining the confidentiality of patients' medical data that include highly sensitive information which make it vulnerable to constant attacks by hackers. An Attack of this kind constitute one of the gravest risks to healthcare systems globally, especially those systems with a vulnerable infrastructure. This makes it easy to target these systems which not only put patients' privacy at risk but also their lives at risk which only serves to stress the importance of the establishment of a strong attack detection system. This study offered a deep learning strategy for identifying data as assault or normal. Data preparation and classification are the two parts of the proposed system. The selected data set has been pre-processed in the first phase then a Convolutional Neural Network (CNN) is employed in the second phase. The Convolutional Neural Network (CNN) model was applied to the dataset in the second phase. The findings imply that the Convolutional Neural Network can identify assaults quicker than previous approaches. The trial results indicated that the data set constructed using the CNN model as the basic classifier had the best classification accuracy of 99.28 with an error rate of 0.0846.