Type-2 fuzzy ontology–aided recommendation systems for IoT–based healthcare

Farman Ali, S. M.Riazul Islam, Daehan Kwak, Pervez Khan, Niamat Ullah, Sang jo Yoo, K. S. Kwak

Research output: Contribution to journalArticlepeer-review

95 Citations (Scopus)

Abstract

The number of people with a chronic disease is rapidly increasing, giving the healthcare industry more challenging problems. To date, there exist several ontology and IoT-based healthcare systems to intelligently supervise the chronic patients for long-term care. The central purposes of these systems are to reduce the volume of manual work in recommendation systems. However, due to the increase of risk and uncertain factors of the diabetes patients, these healthcare systems cannot be utilized to extract precise physiological information about patient. Further, the existing ontology-based approaches cannot extract optimal membership value of risk factors; thus, it provides poor results. In this regards, this paper presents a type-2 fuzzy ontology–aided recommendation systems for IoT-based healthcare to efficiently monitor the patient's body while recommending diets with specific foods and drugs. The proposed system extracts the values of patient risk factors, determines the patient's health condition via wearable sensors, and then recommends diabetes-specific prescriptions for a smart medicine box and food for a smart refrigerator. The combination of type-2 Fuzzy Logic (T2FL) and the fuzzy ontology significantly increases the prediction accuracy of a patient's condition and the precision rate for drug and food recommendations. Information about the patient's disease history, foods consumed, and drugs prescribed is designed in the ontology to deliver decision-making knowledge using Protégé Web Ontology Language (OWL)-2 tools. Semantic Web Rule Language (SWRL) rules and fuzzy logic are employed to automate the recommendation process. Moreover, Description Logic (DL) and Simple Protocol and RDF Query Language (SPARQL) queries are used to evaluate the ontology. The experimental results show that the proposed system is efficient for patient risk factors extraction and diabetes prescriptions.

Original languageEnglish
Pages (from-to)138-155
Number of pages18
JournalComputer Communications
Volume119
Early online date15 Mar 2018
DOIs
Publication statusPublished - 1 Apr 2018
Externally publishedYes

Cite this