An optimized Bi-LSTM with deep learning-based intrusion detection system in healthcare using Blockchain
Swathi Darla, C. Naveena, B N Ajay, Muzameel Ahmed · Information Security Journal A Global Perspective · 2025
Healthcare service quality has significantly improved due to its integration with the Internet of Things (IoT). Cloud network-based servers provide storage, interaction, and problem-solving facilities here. Nevertheless, these also face cyber security threats and issues. So, Intrusion detection systems (IDS), which enable the detection of a wide range of hostile attempts against network security, are becoming a vital tool in healthcare services now. IDS protect patient health records and medical communications against threats in the network layer. This paper suggests a Deep Blockchain-based IDS (DBC-IDS) for ensuring data security and privacy in cloud networks used by healthcare services. A Hybrid Deep Neural Network with Optimized Bidirectional Long Short-Term Memory (HDNN OBi-LSTM) employs the IDS to identify attacks during network data transfer in cloud systems. To reduce the classification time, a feature vector selection technique based on Adaptive Gray Wolf Optimization (AGWO) generates a collection of feature vectors. Golden Search Optimization is then applied to select the best weights for the hidden layers and elevate the classifier’s Bidirectional Long Short-Term Memory (Bi-LSTM) classifier detection rate. The suggested method achieved an accuracy of 98.25%, demonstrating its superior performance and increased efficiency compared to the existing techniques.