Annonce en lien avec l’Action/le Réseau : aucun
Laboratoire/Entreprise : Laboratoire Hubert Curien
Durée : 4 – 6 months
Contact : christine.largeron@univ-st-etienne.fr
Date limite de publication : 2019-02-11
Contexte :
This project aims to understand how people’s stance on contentious issues is influenced by the information diffusions they are exposed to. Our incipient research concerning the discussion on Reddit around Brexit indicates that users change their attitude depending on what content they are exposed to. In this project, we intend to uncover the precise interaction patterns that are most effective in swaying user attitude, and accurately predict what would be a users’ opinion in the close future, based on their past actions and the discussion they were exposed to recently.
Sujet :
This project proposes to jointly use social community detection developed at the University Jean Monnet [7,8], and information diffusions analysis tools [9] currently under development at the University of Technology Sydney to analyse user activity within particular discussion topics together with their stance. More precisely, in this project we will develop upon recently proposed information diffusion mechanisms [5,6] to build spatio-temporal modelling tools, which allow understanding the diffusion paths of information through particular online communities. This project will advance both the theoretical knowledge in the spatio-temporal modelling information diffusion and detecting online communities, and also will build an online platform for visualising in real time the spread of particular discussion topics via popular platforms (e.g. Twitter). The students in this project will employ two readily available datasets around Brexit: a Reddit dataset and a Twitter dataset.
Opportunities. The work in this project will follow upon incipient analysis on the same datasets, and profiling measure which show the feasibility of the approach. The student who will undertake this project will be based in St Etienne, at the Hubert Curien laboratory.
Profil du candidat :
Student in last year of engineering school or Master 2 Computer Science with skills in Data mining / Machine learning
Formation et compétences requises :
Good math skills: probabilities and stochastic calculus; linear algebra
Good programming skills
Background Machine Learning and/or Data Science methods;
Performing (computer) experiments and analysing results
Git, R/Python, desire to make sense of real data and solve real issues
Adresse d’emploi :
Laboratoire Hubert Curien
Université Jean Monnet – Saint-Etienne
18 rue Lauras
42000 Saint-Etienne
Document attaché : 2019_10_02_M2_Largeron_Rizoiu_2019-2020.pdf