Narrow Results

Within the context of agent-based Monte-Carlo simulations, we study the problem of the fluctuations of tax evasion in a community of honest citizens and tax evaders by using the version of the nonequilibrium Zaklan model proposed by Lima (2010). The studied evolutionary dynamics of tax evasion are driven by a non-equilibrium majority-vote model of M. J. Oliveira, with the objective to attempt to control the fluctuations of the tax evasion in the observed community in which citizens are localized on the nodes of the Stauffer–Hohnisch–Pittnauer networks.

We study a nonequilibrium model with up–down symmetry and a noise parameter q known as majority-vote model (MVM) of Oliveira 1992 on opinion-dependent network or Stauffer–Hohnisch–Pittnauer (SHP) networks. By Monte Carlo (MC) simulations and finite-size scaling relations the critical exponents β∕ν, γ∕ν and 1∕ν and points q_c and U* are obtained. After extensive simulations, we obtain β∕ν = 0.230(3), γ∕ν = 0.535(2) and 1∕ν = 0.475(8). The calculated values of the critical noise parameter and Binder cumulant are q_c = 0.166(3) and U* = 0.288(3). Within the error bars, the exponents obey the relation 2β∕ν + γ∕ν = 1 and the results presented here demonstrate that the MVM belongs to a different universality class than the equilibrium Ising model on SHP networks, but to the same class as majority-vote models on some other networks.

We study a nonequilibrium model with up-down symmetry and a noise parameter q known as majority-vote model (MVM) of [M. J. Oliveira, J. Stat. Phys.66, 273 (1992)] with heterogeneous agents on square lattice (SL). By Monte Carlo (MC) simulations and finite-size scaling relations, the critical exponents β∕ν, γ∕ν and 1∕ν and points q_c and U* are obtained. After extensive simulations, we obtain β∕ν = 0.35(1), γ∕ν = 1.23(8) and 1∕ν = 1.05(5). The calculated values of the critical noise parameter and Binder cumulant are q_c = 0.1589(4) and U* = 0.604(7). Within the error bars, the exponents obey the relation 2β∕ν + γ∕ν = 2 and the results presented here demonstrate that the MVM heterogeneous agents belongs to a different universality class than the nonequilibrium MVM with homogeneous agents on SL.

In this paper, we use the version of the nonequilibrium Zaklan model via agent-based Monte-Carlo simulations to study the problem of the fluctuations of the tax evasion on a heterogeneous agents community of honest and tax evaders citizens. The time evolution of this system is performed by a nonequilibrium model known as majority-vote model, but with a different probability for each agent to disobey the majority vote of its neighbors.

In this work, we use Monte-Carlo simulations to study the control of the fluctuations for tax evasion in the economics model proposed by [G. Zaklan, F. Westerhoff and D. Stauffer, J. Econ. Interact. Coordination. 4 (2009) 1; G. Zaklam, F.W.S. Lima and F. Westerhofd, Physica A387 (2008) 5857.] via a nonequilibrium model with two states ($-1,+1$) and a noise $q$ proposed for [M. J. Oliveira, J. Stat. Phys.66 (1992) 273] and known as Majority-Vote model (MVM) and Sánchez–López-Rodríguez model on communities of agents or persons on some topologies as directed and undirected Barabási–Albert networks and Erdös–Rényi random graphs, Apollonian networks, directed small-world networks and Stauffer–Hohnisch–Pittnauer networks. The MVM is applied around the noise critical $q_c$ to evolve the Zaklan model.

Methods for combining multiple classifiers have been developed for improved performance in pattern recognition. This paper examines nine correlated classifiers from the perspective of majority voting. It demonstrates that relationships between the classifiers can be observed from the voting results, that the error reduction ability of a combination varies inversely with the correlation among the classifiers to be combined, and that the correlation coefficient is an effective measure for selecting a subset of classifiers for combination to achieve the best results.

In order to improve recognition results, decisions of several classifiers can be combined. The combination can be accomplished in different ways depending on the types of information produced by the individual classifiers. This chapter considers combination methods that can be applied when the information is provided at both the abstract and measurement levels.

For abstract-level classifiers, the combination methods discussed in this chapter consist of majority vote, weighted majority vote with weights derived from a genetic search algorithm, Bayesian formulation, and Behavior-Knowledge Space method. To combine the decisions of measurement-level classifiers, a multi-layer perceptron is used.

Theoretical considerations and experimental results on handwritten characters are presented, and the results show that combining multiple classifiers is an effective means of producing highly reliable decisions for both categories of classifiers.