The Ground Truth Project
Harnessing novel machine learning to see if social media can be leveraged detect those prone to mental illness, suicidal ideation or people at risk of relapse.
By 2030, depression is forecast to be the second largest cause of disease burden worldwide. Disability arising from depression costs the Australian economy a reported $14.9 billion annually, with more than 6 million working days lost each year. Many people experiencing depression do not seek help. A major advance in the field would result if we could accurately identify risk in a timely way.
A new, potentially game-changing idea is that data from sources that individuals generate in the normal course of their lives, can provide the basis by which risk can be detected. One such data source is social media. Social media may provide an extraordinary opportunity to identify individuals at risk of developing depression, anxiety and suicide or those at risk of relapse, using data that individuals generate ‘naturally’.
In this study, we focus specifically on one type of social media, blogging. We profile individuals' mental health by harnessing novel machine learning based analyses of their blog posts. By doing this, we may be able to determine new markers of mental health.
History of the project
Social media is of interest to researchers at the Institute because large numbers of people in Australia and worldwide, use these platforms. In addition, social media enables passive data collection, that occurs in real-time.
The current project evolved from past studies in the area which have examined the markers of mental health from the social media platform Twitter, including our own We Feel project and the Suicidality on Twitter Project.Back to top
The project was completed in June 2017 and results are currently being prepared for publication.
Outcomes are currently being prepared for publication.
NHMRC John Cade Fellowship in Mental Health Research