Detecting Depression Signs on Social Media: A Systematic Literature Review.

Area of Science:

• Computational Social Science
• Mental Health Informatics
• Digital Psychiatry

Background:

• Depression is a severe mental health condition with significant public health implications.
• Social media platforms generate vast amounts of user data that may contain indicators of mental health status.
• Early detection of depression signs on social media can aid public healthcare efforts.

Purpose of the Study:

• To systematically review and synthesize primary research on depression sign detection using social media data.
• To identify trends, common methodologies, and tools used in this research area between 2016 and mid-2021.

Main Methods:

• A comprehensive literature search was conducted across five major digital libraries and Google Scholar.
• Thirty-four primary studies published between 2016 and mid-2021 were selected and analyzed.
• Data extraction and synthesis focused on social media platforms, feature extraction techniques, machine learning algorithms, and statistical analysis methods.

Main Results:

• Twitter emerged as the most frequently studied social media platform for depression detection.
• Word embedding was the predominant method for linguistic feature extraction.
• Support Vector Machine (SVM) was the most utilized machine learning algorithm, with Python libraries being the most popular computing tools.
• Cross-validation (CV) was the most common statistical analysis technique.

Conclusions:

• Social media data, combined with computational tools and classification methods, offers a valuable resource for detecting depression signs.
• The findings provide insights into the current landscape of social media-based depression detection research.
• This approach supports public healthcare initiatives for early identification and intervention of depression.