Prognosis of Vitamin D Deficiency Severity using SMOTE optimized Machine Learning Models

Abstract

According to global health studies, Deficiency in vitamin D (VDD) is a major public health problem, and there is a strong need for developing a prediction method that can use non-invasive approaches. Invasive methods include the use of medical instruments and tests that can take a long time to predict the outcome of a VDD procedure. This paper proposes to use machine learning classification algorithms for predicting VDD. The machine learning algorithms include Random Forest (RF), Multi-Layer Perceptron (MLP), K-Nearest Neighbour (KNN), Support Vector Machine (SVM), Decision Tree (DT), Gradient Boosting (GB), Stochastic Gradient Descent (SGD), AdaBoost classifier (AB), Extra Trees classifier algorithm (ET), and Logistic Regression (LR). The article evaluates the output of different machine learning classification methods for estimating the severity of VDD in humans. The main goal is to find the most reliable model and incorporate it into the prediction system.