Researching Mental Health Disorders in the Era of Social Media: Systematic Review

Abstract

BackgroundMental illness is quickly becoming one of the most prevalent public health problems worldwide. Social network platforms, where users can express their emotions, feelings, and thoughts, are a valuable source of data for researching mental health, and techniques based on machine learning are increasingly used for this purpose.ObjectiveThe objective of this review was to explore the scope and limits of cutting-edge techniques that researchers are using for predictive analytics in mental health and to review associated issues, such as ethical concerns, in this area of research.MethodsWe performed a systematic literature review in March 2017, using keywords to search articles on data mining of social network data in the context of common mental health disorders, published between 2010 and March 8, 2017 in medical and computer science journals.ResultsThe initial search returned a total of 5386 articles. Following a careful analysis of the titles, abstracts, and main texts, we selected 48 articles for review. We coded the articles according to key characteristics, techniques used for data collection, data preprocessing, feature extraction, feature selection, model construction, and model verification. The most common analytical method was text analysis, with several studies using different flavors of image analysis and social interaction graph analysis.ConclusionsDespite an increasing number of studies investigating mental health issues using social network data, some common problems persist. Assembling large, high-quality datasets of social media users with mental disorder is problematic, not only due to biases associated with the collection methods, but also with regard to managing consent and selecting appropriate analytics techniques.

Trends of infodemiology studies: a scoping review

“…The data collection techniques used in the studies included in this review consisted of aggregating data from Web 1.0, Web 2.0 technologies or epidemiological data. This is in confirmation of Wongkoblap, Vadillo and Curcin (); however, they directly collected data from Facebook participants. One reason for this might be that getting data from Facebook requires users’ consent and Facebook does not allow access to interactions between users.…”

Anxious Depression Prediction in Real-time Social Data

“…Various secondary studies have also given insights to the models built for identifying depression on social media like Facebook, Twitter, and web forums (Guntuku et al, 2017;Seabrook, 2016;Wongkoblap, 2017).…”

High-resolution Temporal Representations of Alcohol and Tobacco Behaviors from Social Media Data

“…We used geo-located data (polygon) from the Twitter API, and restricted to data only from USA, resulting in approximately 50 million Tweets per month. Stop words, Retweets and URLs were removed, and all characters converted to lower case per convention in Twitter data processing [26, 88]. A few intervals of hours in the 3rd week of September had missing data (considered missing if there were fewer than 5000 Tweets in an hour).…”