Valentina Pansanella, Alina Sirbu, Janos Kertesz, Giulio Rossetti [email protected]

People increasingly shape their opinions by accessing and discussing content shared on social networking websites. These platforms contain a mixture of other users' shared opinions and content from mainstream media sources. While online social networks have fostered information access and diffusion, they also represent optimal environments for the proliferation of polluted information and contents, which are argued to be among the co-causes of polarization/radicalization phenomena. Moreover, recommendation algorithms - intended to enhance platform usage - likely augment such phenomena, generating the so-called Algorithmic Bias. In this work, we study the effects of the combination of social influence and mass media influence on the dynamics of opinion evolution in a biased online environment, using a recent bounded confidence opinion dynamics model with algorithmic bias (Sirbu et al., 2019) as a baseline and adding the possibility to interact with one or more media outlets, modeled as stubborn agents. We analyzed four different media landscapes and found that an open-minded population is more easily manipulated by external propaganda - moderate or extremist - while remaining undecided in a more balanced information environment. By reinforcing users' biases, recommender systems appear to help avoid the complete manipulation of the population by external propaganda.

After installig NDlib Python Library

pip install ndlib

In the code folder you can find simulation_example.ipynb with the pipeline to reproduce this work.

Import libraries

import ndlib.models.ModelConfig as mc
import ndlib.models.opinions as op
import networkx as nx

You need to create a graph representing your simulation population

graph = nx.complete_graph(250)

Set up model parameters

mu = 0.5
epsilon = 0.3
gamma = 1.0
gamma_media = gamma
pm = 0.5
k = 3
media_op = [0.05, 0.5, 0.95]

Set up simulation parameters

max_iterations = 1000
sensibility = 0.00001
nsteady = 1000
nodeStatus = True
progressBar = True
drop_evolution = False

Create model

model = op.AlgorithmicBiasMediaModel(graph)

Create configuration

config = mc.Configuration()
config.add_model_parameter("mu", mu)
config.add_model_parameter("epsilon", epsilon)
config.add_model_parameter("gamma", gamma)
config.add_model_parameter("gamma_media", gamma)
config.add_model_parameter("p", pm)
config.add_model_parameter("k", k)

Configure model

model.set_initial_status(config)
model.set_media_opinions(media_op)

Perform iterations untill convergence (if possible)

iterations = model.steady_state(max_iterations=max_iterations, nsteady=nsteady, sensibility=sensibility, node_status=nodeStatus, progress_bar=progressBar, drop_evolution=drop_evolution)