Suivi et détection des idéations suicidaires dans les médias sociaux

L’utilisation croissante des medias sociaux permet un acces sans precedent aux comportements, aux pensees et aux sentiments des individus. Nous nous interessons ici a l’evolution des etats emotionnels des individus captes au travers des services de microblogging de type Twitter. Notre objectif est de predire l’apparition d’ideations suicidaires. Dans ce travail, nous avons mis en place une chaine de traitements permettant d’extraire des caracteristiques a partir des messages refletant l’etat emotionnel. Puis, nous appliquons un modele base sur les Conditionnal Random Fields pour predire un nouvel etat. L’originalite de l’approche est de prendre en compte l’historique des etats emotionnels pour predire le nouvel etat. Une experimentation preliminaire nous a permis d’evaluer notre approche sur des cas reels d’utilisateurs de Twitter. Ces type d’approche permet de mieux comprendre les liens entre expressions dans les medias sociaux et ideations suicidaires ainsi que les transitions entre etats emotionnels.

[1]  D. Lester,et al.  Twitter postings and suicide: An analysis of the postings of a fatal suicide in the 24 hours prior to death , 2015 .

[2]  John Pestian,et al.  Using Natural Language Processing to Classify Suicide Notes , 2008, BioNLP.

[3]  Sandra Bringay,et al.  Concept drift vs suicide: comment l'un peut prévenir l'autre? , 2016, EGC.

[4]  J. Russell,et al.  A 12-Point Circumplex Structure of Core Affect. , 2011, Emotion.

[5]  Eric Horvitz,et al.  Predicting postpartum changes in emotion and behavior via social media , 2013, CHI.

[6]  H. Sueki,et al.  The association of suicide-related Twitter use with suicidal behaviour: a cross-sectional study of young internet users in Japan. , 2015, Journal of affective disorders.

[7]  P. Burnap,et al.  A Naïve Bayes Approach to Classifying Topics in Suicide Notes , 2012, Biomedical informatics insights.

[8]  Andrew McCallum,et al.  An Introduction to Conditional Random Fields , 2010, Found. Trends Mach. Learn..

[9]  Pete Burnap,et al.  Machine Classification and Analysis of Suicide-Related Communication on Twitter , 2015, HT.

[10]  L. Flashman,et al.  Predicting the Risk of Suicide by Analyzing the Text of Clinical Notes , 2014, PloS one.

[11]  Francis R. Bach,et al.  Online Learning for Latent Dirichlet Allocation , 2010, NIPS.

[12]  R. Larsen,et al.  Promises and problems with the circumplex model of emotion. , 1992 .

[13]  Irfan A. Essa,et al.  Beyond Sentiment: The Manifold of Human Emotions , 2012, AISTATS.

[14]  Elizabeth D. Cox,et al.  Feeling bad on Facebook: depression disclosures by college students on a social networking site , 2011, Depression and anxiety.

[15]  Cecilia Ovesdotter Alm,et al.  Detecting Distressed and Non-distressed Affect States in Short Forum Texts , 2012 .

[16]  Thomas Wetter,et al.  Screening Internet forum participants for depression symptoms by assembling and enhancing multiple NLP methods , 2015, Comput. Methods Programs Biomed..

[17]  Abby D. Adler,et al.  A Mixed Methods Approach to Identify Cognitive Warning Signs for Suicide Attempts , 2016, Archives of suicide research : official journal of the International Academy for Suicide Research.

[18]  Andrew McCallum,et al.  Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data , 2001, ICML.

[19]  Pete Burnap,et al.  Analysing the connectivity and communication of suicidal users on twitter , 2016, Comput. Commun..

[20]  Junlan Feng,et al.  Robust Sentiment Detection on Twitter from Biased and Noisy Data , 2010, COLING.