Predicting the level of generalized anxiety disorder of the coronavirus pandemic among college age students using artificial intelligence technology

Abstract

Introduction: Emerging reports indicate heightened anxiety among university students during the Corona pandemic. Implications of which can impact their academic performance. Artificial intelligence (AI) through machine learning can be used to predict which students are more susceptible to anxiety which can inform closer monitoring and early intervention. To date, there are no studies that have explored the efficacy of AI to predict anxiety among college students. Objective: to develop the best fit model to predict anxiety and to rank the most important factors affecting anxiety. Method: Data was collected using an online survey that included general information; Covid-19 stressors and (GAD-7). This scale categorizes level of anxiety to none, mild, moderate, and severe. We received 917 survey answers. Several machine learning classifiers were used to develop the best fit model to predict student level of anxiety. Results: the best performance based on AUC is AdaBoost (0.943) followed by neural network (0.936). Highest accuracy and F1 were for neural network (0.754) and (0.749) respectively, then neural network selected to be the best fit model. The three scoring methods revealed that the top three features that predicted anxiety to be gender; sufficient support from family and friends; and fixed family income. Conclusion: Neural network model can assist college counselors to predict which students are going through anxiety and revealed the top three features for heightened student anxiety to be gender, a support system, and family fixed income. This information can alter college councilors for early mental intervention.