A Machine Learning Study to Predict Anxiety on Campuses in Lebanon

Area of Science:

• Mental Health Research
• Machine Learning Applications
• Higher Education Studies

Background:

• Global university students are experiencing a significant mental health crisis.
• The COVID-19 pandemic has intensified mental health challenges among students.
• Understanding and predicting student anxiety is crucial for timely intervention.

Purpose of the Study:

• To investigate mental health challenges among university students in Lebanon.
• To develop and evaluate machine learning models for predicting anxiety symptoms.
• To identify key predictors of anxiety in the student population.

Main Methods:

• A survey was conducted among 329 university students in Lebanon.
• Machine learning algorithms (logistic regression, MLP, SVM, RF, XGBoost) were employed.
• Models predicted anxiety symptoms using demographic and self-rated health data.

Main Results:

• The Multi-Layer Perceptron (MLP) model achieved the highest performance with an AUC score of 80.70%.
• Self-rated health emerged as the most significant predictor of anxiety symptoms.
• The study identified key demographic and health-related factors associated with student anxiety.

Conclusions:

• Machine learning, particularly MLP, shows high efficacy in predicting student anxiety.
• Self-rated health is a critical factor in assessing and predicting student mental well-being.
• Further research incorporating data augmentation and multi-class predictions is recommended.