Abstract:This paper investigates how the heterogenous relationships around us affect the spread of diverse opinions in the population. We apply the Potts model, derived from condensed matter physics on signed networks, to multi-opinion propagation in complex systems with logically contradictory interactions. Signed networks have received increasing attention due to their ability to portray both positive and negative associations simultaneously, while the Potts model depicts the coevolution of multiple states affected by interactions. Analyses and experiments on both synthetic and real signed networks reveal the impact of the topology structure on the emergence of consensus and the evolution of balance in a system. We find that, regardless of the initial opinion distribution, the proportion and location of negative edges in the signed network determine whether a consensus can be formed. The effect of topology on the critical ratio of negative edges reflects two distinct phenomena: consensus and the multiparty situation. Surprisingly, adding a small number of negative edges leads to a sharp breakdown in consensus under certain circumstances. The community structure contributes to the common view within camps and the confrontation (or alliance) between camps. The importance of inter- or intra-community negative relationships varies depending on the diversity of opinions. The results also show that the dynamic process causes an increase in network structural balance and the emergence of dominant high-order structures. Our findings demonstrate the strong effects of logically contradictory interactions on collective behaviors, and could help control multi-opinion propagation and enhance the system balance.

Potts Model Based on a Markov Process Computation Solves the Community Structure Problem Effectively

Community structure detection based on Potts model and network's spectral characterization

Analysis of Stability of Community Structure Across Multiple Hierarchical Levels

A Novel Dynamics Combination Model Reveals the Hidden Information of Community Structure

Local resolution-limit-free Potts model for community detection

Scalably Revealing the Dynamics of Soft Community Structure in Complex Networks

Fast Accurate Community Detection Based on Dynamics Process and Stability Optimization

A New Dynamic Community Model For Social Networks

Pitman–Yor Process Mixture Model for Community Structure Exploration Considering Latent Interaction Patterns*

Modeling multi-opinion propagation in complex systems with heterogeneous relationships via Potts model on signed networks

A Stochastic Model For Detecting Heterogeneous Link Communities In Complex Networks

A Spin-Glass Model Based Local Community Detection Method in Social Networks

Markov Dynamics based on Link-node-link Random Walk to Reveal Link Communities in Complex Networks

Community structure exploration considering latent link patterns in complex networks

Multiscale behavior of financial time series model from Potts dynamic system

Phase Transition Model for Community Detection

Multiscale Community Estimation Based on Temporary Local Balancing Strategy

Multiresolution community detection for megascale networks by information-based replica correlations

Evolutionary Markov Dynamics for Network Community Detection

Coarse-Grained Monte Carlo Simulations of the Phase Transition of the Potts Model on Weighted Networks