Predicting sun protection measures against skin diseases using machine learning approaches

Area of Science:

• Dermatology and Pharmacology
• Public Health
• Artificial Intelligence

Background:

• Rising skin cancer rates underscore the need for effective solar radiation protection strategies.
• Understanding public sun protection practices is critical in dermatology and pharmacology.
• The Arabian Peninsula presents a unique demographic for studying sun exposure behaviors.

Purpose of the Study:

• To analyze public sun protection behaviors in Arabian Peninsula regions.
• To identify key factors influencing sun protection practices.
• To develop predictive models for public sun protection measures.

Main Methods:

• A simple random survey was employed to gather data on sun protection habits.
• Machine learning algorithms, specifically Artificial Neural Network (ANN) and Support Vector Machine (SVM), were utilized.
• Model performance was rigorously evaluated using confusion matrices and receiver operating characteristic curves.

Main Results:

• Nearly half of respondents (49%) exhibit high levels of sun protection, while 51% demonstrate low levels.
• The Support Vector Machine (SVM) model significantly outperformed the Artificial Neural Network (ANN) in predicting sun protection behaviors.
• Key predictors for sun protection practices include sunburn history, gender, seat belt usage, UV index awareness, income, and physical activity.

Conclusions:

• Identified predictors can inform the development of targeted public health policies.
• Enhancing public awareness regarding the risks of solar radiation is essential for preventing skin diseases.
• Effective policies can mitigate the detrimental effects of sun exposure and reduce skin cancer incidence.