-
Benchmarking symbolic regression constant optimization schemes
Authors:
L. G. A dos Reis,
V. L. P. S. Caminha,
T. J. P. Penna
Abstract:
Symbolic regression is a machine learning technique, and it has seen many advancements in recent years, especially in genetic programming approaches (GPSR). Furthermore, it has been known for many years that constant optimization of parameters, during the evolutionary search, greatly increases GPSR performance However, different authors approach such tasks differently and no consensus exists regar…
▽ More
Symbolic regression is a machine learning technique, and it has seen many advancements in recent years, especially in genetic programming approaches (GPSR). Furthermore, it has been known for many years that constant optimization of parameters, during the evolutionary search, greatly increases GPSR performance However, different authors approach such tasks differently and no consensus exists regarding which methods perform best. In this work, we evaluate eight different parameter optimization methods, applied during evolutionary search, over ten known benchmark problems, in two different scenarios. We also propose using an under-explored metric called Tree Edit Distance (TED), aiming to identify symbolic accuracy. In conjunction with classical error measures, we develop a combined analysis of model performance in symbolic regression. We then show that different constant optimization methods perform better in certain scenarios and that there is no overall best choice for every problem. Finally, we discuss how common metric decisions may be biased and appear to generate better models in comparison.
△ Less
Submitted 2 December, 2024;
originally announced December 2024.
-
Phase transitions and universality in the Sznajd model with anticonformity
Authors:
Matheus Calvelli,
Nuno Crokidakis,
Thadeu J. P. Penna
Abstract:
In this work we study the dynamics of opinion formation in the Sznajd model with anticonformity on regular lattices in two and three dimensions. The anticonformity behavior is similar to the introduction of Galam's contrarians in the population. The model was previously studied in fully-connected networks, and it was found an order-disorder transition with the order parameter exponent $β=1/2$ calc…
▽ More
In this work we study the dynamics of opinion formation in the Sznajd model with anticonformity on regular lattices in two and three dimensions. The anticonformity behavior is similar to the introduction of Galam's contrarians in the population. The model was previously studied in fully-connected networks, and it was found an order-disorder transition with the order parameter exponent $β=1/2$ calculated analytically. However, the other phase transition exponents were not estimated, and no discussion about the possible universality of the phase transition was done. Our target in this work is to estimate numerically the other exponents $γ$ and $ν$ for the fully-connected case, as well as the three exponents for the model defined in square and cubic lattices. Our results suggest that the model belongs to the Ising model universality class in the respective dimensions.
△ Less
Submitted 5 September, 2018;
originally announced September 2018.
-
Dynamical behavior of the Niedermayer algorithm applied to Potts models
Authors:
D. Girardi,
T. J. P. Penna,
N. S. Branco
Abstract:
In this work we make a numerical study of the dynamic universality class of the Niedermayer algorithm applied to the two-dimensional Potts model with 2, 3, and 4 states. This algorithm updates clusters of spins and has a free parameter, $E_0$, which controls the size of these clusters, such that $E_0=1$ is the Metropolis algorithm and $E_0=0$ regains the Wolff algorithm, for the Potts model. For…
▽ More
In this work we make a numerical study of the dynamic universality class of the Niedermayer algorithm applied to the two-dimensional Potts model with 2, 3, and 4 states. This algorithm updates clusters of spins and has a free parameter, $E_0$, which controls the size of these clusters, such that $E_0=1$ is the Metropolis algorithm and $E_0=0$ regains the Wolff algorithm, for the Potts model. For $-1<E_0<0$, only clusters of equal spins can be formed: we show that the mean size of the clusters of (possibly) turned spins initially grows with the linear size of the lattice, $L$, but eventually saturates at a given lattice size $\widetilde{L}$, which depends on $E_0$. For $L \geq \widetilde{L}$, the Niedermayer algorithm is in the same dynamic universality class of the Metropolis one, i.e, they have the same dynamic exponent. For $E_0>0$, spins in different states may be added to the cluster but the dynamic behavior is less efficient than for the Wolff algorithm ($E_0=0$). Therefore, our results show that the Wolff algorithm is the best choice for Potts models, when compared to the Niedermayer's generalization.
△ Less
Submitted 19 April, 2012;
originally announced April 2012.
-
Detrended Fluctuation Analysis of Systolic Blood Pressure Control Loop
Authors:
C. E. C. Galhardo,
T. J. P. Penna,
M. Argollo de Menezes,
P. P. S. Soares
Abstract:
We use detrended fluctuation analysis (DFA) to study the dynamics of blood pressure oscillations and its feedback control in rats by analyzing systolic pressure time series before and after a surgical procedure that interrupts its control loop. We found, for each situation, a crossover between two scaling regions characterized by exponents that reflect the nature of the feedback control and its…
▽ More
We use detrended fluctuation analysis (DFA) to study the dynamics of blood pressure oscillations and its feedback control in rats by analyzing systolic pressure time series before and after a surgical procedure that interrupts its control loop. We found, for each situation, a crossover between two scaling regions characterized by exponents that reflect the nature of the feedback control and its range of operation. In addition, we found evidences of adaptation in the dynamics of blood pressure regulation a few days after surgical disruption of its main feedback circuit. Based on the paradigm of antagonistic, bipartite (vagal and sympathetic) action of the central nerve system, we propose a simple model for pressure homeostasis as the balance between two nonlinear opposing forces, successfully reproducing the crossover observed in the DFA of actual pressure signals.
△ Less
Submitted 25 August, 2009;
originally announced August 2009.
-
A Harris-Todaro Agent-Based Model to Rural-Urban Migration
Authors:
Aquino L. Espindola,
Jaylson J. Silveira,
T. J. P. Penna
Abstract:
The Harris-Todaro model of the rural-urban migration process is revisited under an agent-based approach. The migration of the workers is interpreted as a process of social learning by imitation, formalized by a computational model. By simulating this model, we observe a transitional dynamics with continuous growth of the urban fraction of overall population toward an equilibrium. Such an equilib…
▽ More
The Harris-Todaro model of the rural-urban migration process is revisited under an agent-based approach. The migration of the workers is interpreted as a process of social learning by imitation, formalized by a computational model. By simulating this model, we observe a transitional dynamics with continuous growth of the urban fraction of overall population toward an equilibrium. Such an equilibrium is characterized by stabilization of rural-urban expected wages differential (generalized Harris-Todaro equilibrium condition), urban concentration and urban unemployment. These classic results obtained originally by Harris and Todaro are emergent properties of our model.
△ Less
Submitted 27 October, 2005;
originally announced October 2005.
-
Rural-urban migration in d-dimensional lattices
Authors:
Aquino L. Espíndola,
T. J. P. Penna,
Jaylson J. Silveira
Abstract:
The rural-urban migration phenomenon is analyzed by using an agent-based computational model. Agents are placed on lattices which dimensions varying from d=2 up to d=7. The localization of the agents in the lattice define their social neighborhood (rural or urban) not being related to their spatial distribution. The effect of the dimension of lattice is studied by analyzing the variation of the…
▽ More
The rural-urban migration phenomenon is analyzed by using an agent-based computational model. Agents are placed on lattices which dimensions varying from d=2 up to d=7. The localization of the agents in the lattice define their social neighborhood (rural or urban) not being related to their spatial distribution. The effect of the dimension of lattice is studied by analyzing the variation of the main parameters that characterizes the migratory process. The dynamics displays strong effects even for around one million of sites, in higher dimensions (d=6, 7).
△ Less
Submitted 8 June, 2005;
originally announced June 2005.
-
An agent-based model to rural-urban migration analysis
Authors:
Jaylson J. Silveira,
Aquino L. Espindola,
T. J. P. Penna
Abstract:
In this paper we analyze the rural-urban migration phenomena as it is usually observed in economies which are in the early stages of industrialization. The analysis is conducted by means of a statistical mechanics approach which builds a computational agent-based model. Agents are placed on a lattice and the connections among them are described via an Ising like model. Simulations on this comput…
▽ More
In this paper we analyze the rural-urban migration phenomena as it is usually observed in economies which are in the early stages of industrialization. The analysis is conducted by means of a statistical mechanics approach which builds a computational agent-based model. Agents are placed on a lattice and the connections among them are described via an Ising like model. Simulations on this computational model show some emergent properties that are common in developing economies, such as a transitional dynamics characterized by continuous growth of urban population, followed by the equalization of expected wages between rural and urban sectors (Harris-Todaro equilibrium condition), urban concentration and increasing of per capita income.
△ Less
Submitted 2 June, 2005;
originally announced June 2005.
-
Sliding Blocks Revisited: A simulational Study
Authors:
A. R. de Lima,
C. Moukarzel,
T. J. P. Penna
Abstract:
A computational study of sliding blocks on inclined surfaces is presented. Assuming that the friction coefficient $μ$ is a function of position, the probability $P(λ)$ for the block to slide down over a length $λ$ is numerically calculated. Our results are consistent with recent experimental data suggesting a power-law distribution of events over a wide range of displacements when the chute angl…
▽ More
A computational study of sliding blocks on inclined surfaces is presented. Assuming that the friction coefficient $μ$ is a function of position, the probability $P(λ)$ for the block to slide down over a length $λ$ is numerically calculated. Our results are consistent with recent experimental data suggesting a power-law distribution of events over a wide range of displacements when the chute angle is close to the critical one, and suggest that the variation of $μ$ along the surface is responsible for this.
△ Less
Submitted 21 July, 1998;
originally announced July 1998.
