Machine learning in health condition check-up: An approach using Breiman's random forest algorithm. (October 2022)

• Record Type: Journal Article
• Authors: Abd Algani, Yousef Methkal; Ritonga, Mahyudin; Kiran Bala, B.; Al Ansari, Mohammed Saleh; Badr, Malek; Taloba, Ahmed I.
• Is Part Of: Measurement. Volume 23(2022)
• Publication Date: 2022-10
• Subjects: Random forest algorithm -- Bagging -- Classifications -- Machine learning -- Health checking; Detectors -- Periodicals; Measurement -- Periodicals; 530.7
• DOI: 10.1016/j.measen.2022.100406
• Languages: English
• ISSNs: 2665-9174

Abstract

Abstract: Nowadays majority of the college students' physical condition is worrying. They are not physically and also mentally healthy. If so, why? Their selection of foods is not consistent. Thus, they are more likely to suffer from chronic illnesses such as diabetes, hypertension, stress, etc. in the future. Awareness should be created to prevent such diseases before they occur. Physiological parameters measured included Systolic (SBP) and Diastolic (DBP) Blood Pressure, Body mass Index (BMI), Blood Serum Cholesterol (BSC), and percentage of Body Fat (%BF). These parameters are retrieved and classified to check the physical health or predict if any abnormalities are found in the health condition of college students. Therefore, to predict and classify their health status using Breiman's Random Forest (RF) Algorithm is proposed in this paper. Of all the classification methods available, random forests offer the greatest accuracy. Random forest method also handles large data with thousands of variables. When a class is more sparse than further classes in the data it can spontaneously balance the data sets. The outcome shows that the proposed Random Forest algorithm is accurate in predicting and checking the health condition of students. Students' physical condition should be diagnosed through this method. By knowing the healthy body parameters of the students, a physician can know whether they are healthy or not.