Export Publication

Fonseca, A. & Louçã, Jorge (2011). Validation in opinion dynamics modelling: Social Impact Theory and The Presidential Elections Case Study. ECCS11.

A. J. Fonseca and J. M. Louçã,  "Validation in opinion dynamics modelling: Social Impact Theory and The Presidential Elections Case Study", in ECCS11, Vienna, 2011

@null{fonseca2011_1716081375545,
	year = "2011",
	url = "https://ciencia.iscte-iul.pt/publications/validation-in-opinion-dynamics-modelling-social-impact-theory-and-the-presidential-elections-case/41224?lang=en"
}

TY  - GEN
TI  - Validation in opinion dynamics modelling: Social Impact Theory and The Presidential Elections Case Study
T2  - ECCS11
AU  - Fonseca, A.
AU  - Louçã, Jorge
PY  - 2011
CY  - Vienna
UR  - https://ciencia.iscte-iul.pt/publications/validation-in-opinion-dynamics-modelling-social-impact-theory-and-the-presidential-elections-case/41224?lang=en
AB  - Opinion formation modelling often suffers from a lack of empirical data validation. Actually, the research on
social dynamics from the statistical physics perspective has been mainly theoretical. The current availability
of large datasets and the ease by which Internet social data can now be collected makes some validation of
theoretical social models a less difficult task. We propose a novel method for evaluating social dynamics
modelling using on-line data gathering. The method includes three distinct phases: (1) data collection, (2)
parameter adjustment and (3) multi-agent modeling. Specifically, we tested the significance of Social Impact
Theory, originally proposed by Latan´e (1981), for characterizing political opinion formation during electoral
periods. This well known mathematical model was tested using more than 4 millions of tweets collected
from the 1st October to the 21st January 2011, concerning the Portuguese presidential elections occurred in
January 2011. Following the data collection, two distinct on-line communities were inspected: the general
Twitter users community, and the traditional news media that were accessible through Twitter feeds. Two
specific parameters from the Social Impact Theory model were analyzed: persuasiveness and supportiveness. These parameters were adjusted to the empirical data series collected from both Twitter and traditional media. Finally, a multi-agent model was conceived representing the overall population. Media noise, represented by news, was injected in the model. The opinion dynamics, known from pools during the campaign, was simulated by adjusting model parameters. This operation was performed on six separate empirical series respecting the talk about the six electoral candidates. The complete process allowed concluding about the explanatory power of Social Impact Theory, and, on the other side, allowed characterizing opinion dynamics in this specific case study.
ER  -