Machine Learning-Based Performance Analysis for Predicting the Severity of Vitamin D Deficiency

Abstract

There is a squeezing need for harmless strategies to expect the seriousness of vitamin D deficiency (VDD), which is a significant worldwide medical condition. The essential information, which remembered Vitamin D levels for the blood, were assembled from 3044 undergrads between the ages of 18 and 21. VDD was anticipated utilizing age, orientation, level, weight, body mass index (BMI), midriff outline, muscle to fat ratio, bone mass, action, daylight openness, and milk utilization. The target of the review is to look at and examine different ML models for anticipating VDD seriousness. The objectives of our procedure are to estimate utilizing an assortment of refined ML calculations and to evaluate the results utilizing different execution measures, for example, Accuracy, Review, F1-measure, Exactness, and Region under the bend of a beneficiary working trademark a (ROC). The empirical data were checked with the statistical McNemar's test. The last trial results showed that the Arbitrary Timberland Classifier beat the preparation and testing Vitamin D datasets with an exactness of 96%. McNemar's factual tests exhibit that the RF classifier performs better compared to different classifiers.