1 of 4

A machine learning web application for screening social anxiety disorder based on participants’ emotion regulation (ML-SAD)

Sara Ahmadi Majd, Mohamad Rasoul Parsaeian, Mohsen Madani, Hadi Moradi, Abolfazl Mohammadi

Oct 14, IRAIH24

1

2 of 4

Social Anxiety Disorder (SAD) and Emotion Regulation

2

Social Anxiety Disorder (SAD): SAD involves intense, persistent fear of social situations where one may feel embarrassed, humiliated, or negatively judged.

Diagnostic Challenges: Many individuals with SAD go undiagnosed, often due to a lack of awareness and or avoidance of seeking a psychiatrist.

Emotion Regulation and SAD: Individuals with SAD often struggle with emotion regulation, which is the process of managing the type, intensity, and duration of emotions. Different emotion regulation strategies are employed by people with SAD vs. healthy individuals.

Social Anxiety Disorder, or SAD, is a prevalent and often debilitating condition where individuals experience overwhelming fear in social situations. This fear typically stems from worries about embarrassment or negative judgment. However, SAD frequently remains undiagnosed. Patients may not seek help due to social stigma or a lack of awareness about the disorder. Emotion regulation plays a crucial role in this context, as people with SAD struggle with managing their emotional responses, particularly in high-stress situations. They might experience an inability to control the intensity or type of emotions, which makes social interactions even more challenging. By focusing on these aspects, our study aims to develop an approach that leverages emotion regulation patterns to better predict and understand SAD.

3 of 4

Study Design and Methodology

3

Multimedia Scenario-based Assessment

Participants were exposed to social scenarios, rating their use of emotion regulation strategies such as reappraisal, suppression, and avoidance.

Machine Learning Models

Algorithms such as Random Forest were applied to identify patterns and accurately predict SAD.

4 of 4

Key Results and Implications

4

High Model Accuracy

Random Forest achieved accuracies above 85%.

Implications for Diagnosis

Results suggest emotion regulation can serve as a reliable predictor for SAD.

Future Applications

Machine learning models like these could augment traditional diagnostic methods, providing early and accurate SAD detection.

Significant difference

There were significant differences between the SAD group and the normal group.