Vol.:(0123456789) 1 3
Network Modeling Analysis in Health Informatics and Bioinformatics (2020) 9:22
https://doi.org/10.1007/s13721-020-0226-0
ORIGINAL ARTICLE
Multi‑modal social and psycho‑linguistic embedding via recurrent
neural networks to identify depressed users in online forums
Anu Shrestha
1
· Edoardo Serra
1
· Francesca Spezzano
1
Received: 4 January 2020 / Revised: 4 March 2020 / Accepted: 8 March 2020
© Springer-Verlag GmbH Austria, part of Springer Nature 2020
Abstract
Depression is the most common mental illness in the US, with 6.7% of all adults experiencing a major depressive episode.
Unfortunately, depression extends to teens and young users as well and researchers have observed an increasing rate in recent
years (from 8.7% in 2005 to 11.3% in 2014 in adolescents and from 8.8 to 9.6% in young adults), especially among girls and
women. People themselves are a barrier to fghting this disease as they tend to hide their symptoms and do not receive treat-
ments. However, protected by anonymity, they share their sentiments on the Web, looking for help. In this paper, we address
the problem of detecting depressed users in online forums. We analyze user behavior in the ReachOut.com online forum, a
platform providing a supportive environment for young people to discuss their everyday issues, including depression. We
propose an unsupervised technique based on recurrent neural networks and anomaly detection to detect depressed users.
We examine the linguistic style of user posts in combination with network-based features modeling how users connect in
the forum. Our results on detecting depressed users show that both psycho-linguistic features derived from user posts and
network features are good predictors of users facing depression. Moreover, by combining these two sets of features, we can
achieve an F1-measure of 0.64 and perform better than baselines.
Keywords Depression detection · Online forums · Multi-modal user representation · Unsupervised classifcation ·
Recurrent neural networks
1 Introduction
Depression is a mental illness commonly seen in people
(6.7% of all U.S. adults have experienced at least one major
depressive episode), which negatively afects their thoughts
and behaviors. Depression causes mood fuctuations and
impermanent emotional responses to the challenges of
everyday life. Especially when lasting for a while and with
moderate or severe intensity, depression may become a seri-
ous health condition. It can cause the afected person to suf-
fer greatly and perform poorly at work, at school, and in
the family. It has been one of the common problems seen
in tens of millions of people. At its worst, depression can
lead to suicide. Close to 800,000 individuals die due to sui-
cide every year. According to a 2015 report by the World
Health Organization, more than 300 million people are
afected by depression. Unfortunately, depression extends
to teens and young users as well and researchers have
observed an increasing rate in recent years (from 8.7% in
2005 to 11.3% in 2014 in adolescents and from 8.8 to 9.6%
in young adults), especially among girls and women. Very
few people in the world receive the treatments provided for
depression. In many countries, fewer than 10% of people in
need receive such treatments. One of the barriers to this is
the people themselves. They tend to hide their symptoms to
avoid being known as psychiatric patients or because people
are unaware of the condition and what is happening with
them. Online forums and social media are platforms where
This paper is an extended version of the conference paper “Anu
Shrestha and Francesca Spezzano, Detecting Depressed Users in
Online Forums. In Proceedings of the International Symposium
on Network Enabled Health Informatics, Biomedicine and
Bioinformatics (HI-BI-BI 2019)”, in conjunction with ASONAM
2019. Shrestha and Spezzano (2019).
* Francesca Spezzano
francescaspezzano@boisestate.edu
Anu Shrestha
anushrestha@u.boisestate.edu
Edoardo Serra
edoardoserra@boisestate.edu
1
Boise State University, Boise, ID 83702, USA