Logo image
Back
Blockchain-Based Secure Cloud Storage for IoT Healthcare Using BiGRU and BiLSTM Models
Conference proceeding

Blockchain-Based Secure Cloud Storage for IoT Healthcare Using BiGRU and BiLSTM Models

Prachi Arihant Chougule, Sarumathi S, Ravi Kant, Raj Kumar Gupta, U. Pushpa Latha and Anvesh Perada
2025 3rd International Conference on Integrated Circuits and Communication Systems (ICICACS), pp 1-6
21 Feb 2025
DOI: https://doi.org/10.1109/ICICACS65178.2025.10968918

Abstract

Accuracy Bidirectional long short term memory Feature extraction health monitoring internet of things (IoT) Medical services Predictive models recurrent neural network (RNN) Secure storage Security Training Cloud Computing Internet of Things
Machine Learning (ML) algorithms have experienced a significant increase in popularity owing to the digitisation of analogue processes and other technological advancements, like the Internet of Things (IoT). Dependable datasets are crucial for the precise forecasting and resolution abilities of these algorithms, which are indispensable in healthcare, cloud computing, engineering, and finance. Conversely, training datasets are vulnerable to manipulation, thereby skewing the results. To address this issue, blockchain-based approaches have been proposed to enhance the reliability and security of cloud-stored E-Health data generated by the Internet of Things. To improve the accuracy and efficacy of healthcare prediction models, our study focusses on an advanced strategy that integrates feature selection with model training. Optimal input data is refined by feature selection employing entropy and correlation coefficient methodologies. The proposed model, employing BiGRU and BiLSTM, surpasses conventional CNNs and BiRNNs, with an average accuracy of 94.57%. When integrated with the Internet of Things (IoT), blockchain technology ensures the secure storage and transmission of electronic health records (EHRs). This strategy enhances predictive accuracy and safeguards sensitive health data in cloud infrastructures by amalgamating feature selection techniques with robust deep learning models.

Metrics

1 Record Views

Details

Logo image