A novel blockchain-enabled heart disease prediction mechanism using machine learning

Muhammad Tufail, Huru Hasanonva, Ui-Jun Baek, Jee-Tae Park, Myung-Sup Kim (Corresponding Author)

Research output: Journal article publicationJournal articleAcademic researchpeer-review

Abstract

Heart disease is one of the leading causes for death in men and women across the globe. Several characteristics that can be monitored to predict the heart disease in the earlier stage are blood pressure, cholesterol level, blood sugar level and body weight. The technology is revolutionizing the existing healthcare infrastructure. With the inclusion of Internet of Things (IoT), now we can monitor patients remotely, store their data, and process it for further analysis. However, the need is to propose new and advanced secured algorithms for fast processing and efficient detection of events. In this article, a machine learning based Sine Cosine Weighted K-Nearest Neighbour (SCA_WKNN) algorithm is proposed for the heart disease prediction that learns from the data being stored in blockchain. Since the data stored in the blockchain are tamper resistant, it acts as an authentic source for learning data and also as a secure storage environment for patient information. The performance of proposed SCA_WKNN is assessed in comparison with other algorithms in terms of accuracy, precision, recall, F-score, and root mean square error. Our analysis indicates that SCA_WKNN achieves 4.59% and 15.61% maximum accuracy than W K-NN and K-NN, respectively. Also, blockchain-based storage is compared with peer-to-peer storage in terms of latency and throughput. The blockchain-based decentralized storage achieves 25.03% maximum throughput than peer-to-peer storage.
Original languageEnglish
Article number108086
JournalComputers and Electrical Engineering
Volume101
DOIs
Publication statusPublished - Jul 2022

Fingerprint

Dive into the research topics of 'A novel blockchain-enabled heart disease prediction mechanism using machine learning'. Together they form a unique fingerprint.

Cite this