Can insulin resistance be predicted with vitamin B12, ferritin, vitamin D and demographic information using machine learning algorithms?

Autor: Yigit, Meltem, Secgin, Yusuf, Oner, Zulal, Olukman, Ozgur
Předmět:
Zdroj: Medicine Science; Mar2024, Vol. 13 Issue 1, p55-60, 6p
Abstrakt: In this study, we aimed to investigate the relationship between insulin resistance and vitamin D, vitamin B12 and ferritin levels. Between January 2022 and October 2023, 322 patients (133 females, 189 males) between the ages of 3-16 years who were admitted to the Pediatrics outpatient clinics of İzmir Bakırçay University Çiğli Training and Research Hospital were included. Pediatric patients with no known chronic disease and measured vitamin B12, vitamin D, ferritin and HOMAIR levels were retrospectively included in the study. Individuals with HOMA-IR<2.5 were considered as (normal) controls and those with HOMA-IR=2.5 were considered as insulin resistant children. HOMA-IR groups were predicted with an accuracy rate of 0.80 with the Random Forest (RF) algorithm, one of the machine learning algorithms. In addition, the highest contribution of RF to the determination of HOMA-IR with the SHAP analyzer was found to be provided by age, followed by vitamin B12. The results of the study revealed that vitamin B12, vitamin D, ferritin level and age are important factors on insulin resistance. Early vitamin D, vitamin B12 and ferritin replacement is important for the control of metabolic diseases in the future. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index