Roghieh Nooripour - Department of Counseling, Faculty of Education and Psychology, Alzahra University, Tehran, Iran.
Hossein Ilanloo - Department of Counseling, Faculty of Education and Psychology, Kharazmi University, Tehran, Iran.
Mohammad Naveshki - Department of Petroleum, Faculty of Reservoir Engineering Oil and Gas, Sahand University of Technology, Tabriz, Iran.
Saeid Naveshki - Department of Remote Sensing and Geographic Information Systems (GIS), Faculty of Remote Sensing and Geographic Information System Geographic, Kharazmi University, Tehran, Iran.
Mojtaba Amiri Majd - Department of Psychology, Faculty of Humanities, Abhar Branch, Islamic Azad University, Abhar, Iran.
Sverker Sikström - Chair of the Cognitive Division, Department of Psychology, Faculty of Education and Psychology, Lund University, Lund, Sweden.
Farnaz Jafari - Department Of Educational Sciences, Faculty of Education and Psychology, Farhangian University, Shiraz, Iran.

Objective: The coronavirus pandemic has presented a significant challenge and brought about dramatic changes for universities and their students. This study evaluated machine learning algorithms for estimating COVID-۱۹ stress levels among Iranian university students. Methods: We conducted an online survey from May ۱۰th to November ۲۰th, ۲۰۲۱, to determine how Iranian university students responded to the COVID-۱۹ outbreak in Iran. The survey invitations were sent to Iranian university students via e-mail, forums, and social media platforms, such as internet advertisements. We collected data from ۳۴۹۰ university students, using sociodemographic characteristics and the COVID-۱۹ Stress Scale (CSS; Nooripour et al. [۲۰۲۲]). The adaptive neuro-fuzzy inference system (ANFIS) network for prediction and fuzzy logic-based rules were used for analyzing the data. For classification, eight machine learning algorithms were employed: support vector machine (SVM), K-nearest neighbors (KNN), random forest, multilayer perceptron, decision tree, and passive-aggressive algorithm. These algorithms were selected based on their principles and suitability for stress detection in the desired category. Results: Among the algorithms, the decision tree algorithm showed the best performance in accurately classifying the data into the correct stress intensity categories. Moreover, analyses revealed that gender, age group, and education significantly influenced stress intensity levels, with men experiencing less stress; stress intensity decreased with age, and higher education was associated with lower stress levels. The results indicated that education and marital status were the most influential parameters for all three top-performing algorithms (random forest, multi-layer perceptron, and decision tree). Conclusion: Our research suggests that innovative methods such as machine learning algorithms can be used to evaluate psychological distress caused by the COVID-۱۹ outbreak, such as stress. Evaluating stress levels can help prevent mental health problems and enhance students’ coping capabilities.

COVID-۱۹, Machine learning, Psychological distress, Student, Stress