Blockchain-Enabled Thyroid Detection Using Voting Classifier and Deep Convolutional Neural Network

Thyroid disorders represent a major global health issue, requiring timely diagnosis and precise classification to ensure effective treatment. Conventional diagnostic techniques, such as fine-needle aspiration cytology (FNAC), frequently yield ambiguous results, emphasizing the necessity for advanced computational solutions. This study proposed the voting classifier based on selecting top base classifiers from lazy classifier and also proposed deep learning model to detect and classify the thyroid conditions over the thyroid dataset. Measure the performance of the proposed models using various evaluation metrics. It shows that proposed voting (soft) ensemble model archives the accuracy of 97.20 %, and comparatively deep CNN model achieves the 94.80 % The proposed ensemble model outperforming traditional classification methods. Additionally, proposed a blockchain based security model to improve the confidentiality and integrity of patient records based on encryption mechanisms and rolebased access controls using smart contracts for the cloud-based system. The results confirm the potential of combining ensemble learning, deep learning, and blockchain to improve the reliability and efficiency of thyroid disease diagnosis.