Our project
An increasing number of users are exploiting online social services to voice thoughts of depression and suicide. Although talking about "suicide" is at the top of the American Association of Suicidology’s list of suicide warning signs, it remains very tough to assess urgency when a suicide ideation is expressed on social media. In this project, we focus on the detection and prevention of suicide in social networks. The DontDoIt project is a result of collaboration between the University of Montpellier and the Psychiatric Emergencies and Post-Acute Care of Montpellier (CHU de Montpellier).Currently, we are analyzing and monitoring the data of patients who attempted suicide, in order to early detect and prevent future suicide attempts.
We have developed some prototypes to collect data from some social media websites such as: Twitter, Facebook, Youtube and Instagram, and created a dashborad to allow psychiatrists to monitor and follow the patients.
Data scraping
Tool used to collect data from different social media.
Visualization
Psychiatrists dashboard to monitor patients online.
Prediction
Alarms predicted by the models. Results are assessed by the psychiatrists.
Publications
A Context-Aware Based Framework to Track Depression and Suicide Ideation on Social Media.
Bilel Moulahi, Jérôme Azé, Sandra Bringay. In The 18th International Conference on Web Information Systems Engineering, WISE 2017, 7-11 Octobre, Russie, 2017
Suivi et détection des idéations suicidaires dans les médias sociaux.
Bilel Moulahi, Jérôme Azé, Sandra Bringay. 28es Journées francophones d’Ingénierie des Connaissances (IC’2017), 2-7 Juillet, Caen, France, 2017
Analysis and Detection of Suicide-Related Posts in Social Data Streams.
Mia Vioulès, Bilel Moulahi, Jérôme Azé, Sandra Bringay. In IBM Journal of Research and Development. 2017.
Concept drift vs suicide: comment l'un peut prévenir l'autre?
Cédric Maigrot, Sandra Bringay, Jérôme Azé. EGC 2016: 219-230
Mining Twitter for Suicide Prevention.
Amayas Abboute, Yasser Boudjeriou, Gilles Entringer, Jérôme Azé, Sandra Bringay, Pascal Poncelet: NLDB 2014: 250-253
