Predicting vitamin D deficiency using optimized random forest classifier.

Area of Science:

• Nutritional Science
• Medical Informatics
• Public Health

Background:

• Vitamin D is crucial for health, obtainable from diet and sunlight.
• Vitamin D deficiency is a growing global health concern.
• Understanding risk factors is vital for effective prevention and management.

Purpose of the Study:

• To predict Vitamin D deficiency using environmental and nutritional factors.
• To employ an optimized random forest (OptRF) classifier for enhanced prediction accuracy.
• To identify significant predictors of Vitamin D deficiency in the Tabuk population.

Main Methods:

• A predictive, cross-sectional, correlational study involving 350 participants in Saudi Arabia.
• Utilized the Weka machine-learning tool with an optimized random forest (OptRF) algorithm.
• Applied advanced feature selection and hyperparameter tuning for model optimization.

Main Results:

• The OptRF model achieved a high accuracy of 91.42% in classifying Vitamin D deficiency.
• Identified significant predictors including female gender, low exercise, and inadequate Vitamin D and Calcium intake.
• Found that age, income, smoking, and sun exposure were less significant predictors.

Conclusions:

• Tabuk citizens, particularly females, are at high risk of Vitamin D deficiency.
• Low physical activity and poor dietary habits are key contributors.
• Emphasizes the importance of screening, community awareness, and potential supplementation to combat Vitamin D deficiency and its associated chronic illnesses.