Comparison of a genetic algorithm and simulated annealing for automatic neural network ensemble development

Soares, Symone; Antunes, Carlos Henggeler; Araújo, Rui

doi:10.1016/j.neucom.2013.05.024

Cited by 50 publications

(26 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are several methods to apply the overproduce-and-choose approach (Soares et al, 2013;Wang and Alhamdoosh, 2013;Yao and Xu, 2006;Zhou et al, 2002), but to our knowledge those sophisticated methods have not been coupled to a variable selection procedure. Consequently, we applied a step-forward selection of the MLP candidates which has been proved to perform similarly to more complex algorithms (Muñoz-Mas et al, 2014).…”

Section: Multilayer Perceptron Ensemble Developmentmentioning

confidence: 99%

Can multilayer perceptron ensembles model the ecological niche of freshwater fish species?

Muñoz‐Mas

Martínez‐Capel

Alcaraz‐Hernández

et al. 2015

Ecological Modelling

View full text Add to dashboard Cite

a b s t r a c tThe potential of Multilayer Perceptron (MLP) Ensembles to explore the ecology of freshwater fish species was tested by applying the technique to redfin barbel (Barbus haasi Mertens, 1925), an endemic and montane species that inhabits the North-East quadrant of the Iberian Peninsula. Two different MLP Ensembles were developed. The physical habitat model considered only abiotic variables, whereas the biotic model also included the density of the accompanying fish species and several invertebrate predictors. The results showed that MLP Ensembles may outperform single MLPs. Moreover, active selection of MLP candidates to create an optimal subset of MLPs can further improve model performance. The physical habitat model confirmed the redfin barbel preference for middle-to-upper river segments whereas the importance of depth confirms that redfin barbel prefers pool-type habitats. Although the biotic model showed higher uncertainty, it suggested that redfin barbel, European eel and the considered cyprinid species have similar habitat requirements. Due to its high predictive performance and its ability to deal with model uncertainty, the MLP Ensemble is a promising tool for ecological modelling or habitat suitability prediction in environmental flow assessment.

show abstract

Section: Multilayer Perceptron Ensemble Developmentmentioning

confidence: 99%

Can multilayer perceptron ensembles model the ecological niche of freshwater fish species?

Muñoz‐Mas

Martínez‐Capel

Alcaraz‐Hernández

et al. 2015

Ecological Modelling

View full text Add to dashboard Cite

show abstract

“…The fitness stretch of Hybrid Simulated Annealing Algorithm has obvious effect. In the initial stage of Genetic Algorithm, individuals with consistent fitness had the similar probability of offspring generation; in the later period, because of significant decrease of temperature [17], the stretch effect got enhanced, so the fitness difference among different individuals is magnified which has made the excellent individuals more superior [18].…”

Section: Modified Simulated Annealing Genetic Algorithmmentioning

confidence: 99%

Optimization of Multiple Traveling Salesman Problem Based on Simulated Annealing Genetic Algorithm

Sheng

Guo

2017

MATEC Web Conf.

View full text Add to dashboard Cite

Abstract. It is very effective to solve the multi variable optimization problem by using hierarchical genetic algorithm. This thesis analyzes both advantages and disadvantages of hierarchical genetic algorithm and puts forward an improved simulated annealing genetic algorithm. The new algorithm is applied to solve the multiple traveling salesman problem, which can improve the performance of the solution. First, it improves the design of chromosomes hierarchical structure in terms of redundant hierarchical algorithm, and it suggests a suffix design of chromosomes; Second, concerning to some premature problems of genetic algorithm, it proposes a self-identify crossover operator and mutation; Third, when it comes to the problem of weak ability of local search of genetic algorithm, it stretches the fitness by mixing genetic algorithm with simulated annealing algorithm. Forth, it emulates the problems of N traveling salesmen and M cities so as to verify its feasibility. The simulation and calculation shows that this improved algorithm can be quickly converged to a best global solution, which means the algorithm is encouraging in practical uses.

show abstract

“…However, in systems where the parameters of all the models are adjusted to the new concept, this adaptation may produce redundant models and affect the system's performance in recurring drifts. Ensembles can employ an ensemble pruning strategy, which is based on pruning the ensemble by selecting a subset of models from the original set of models, and excluding those models that are either detrimental to the ensemble's performance or contain redundant information, normally with the goal of limiting the number of models and/or improving the ensemble accuracy (Soares et al, 2013;Zhang and Chau, 2009).…”

Section: Related Workmentioning

confidence: 99%

An on-line weighted ensemble of regressor models to handle concept drifts

Soares

Araújo

2015

Engineering Applications of Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

a b s t r a c tMany estimation, prediction, and learning applications have a dynamic nature. One of the most important challenges in machine learning is dealing with concept changes. Underlying changes may make the model designed on old data, inconsistent with new data. Also, algorithms usually specialize in one type of change. Other challenge is reusing previously acquired information in scenarios where changes may recur. This strategy improves the learning accuracy and reduces the processing time. Unfortunately, most existing learning algorithms to deal with changes are adapted on a batch basis. This process usually requires a long time, and such data may not reflect the current state of the system. However, even the system is adapted on a sample basis, existing algorithms may adapt slowly to changes and cannot conciliate old and new information. This paper proposes an On-line Weighted Ensemble (OWE) of regressor models which is able to learn incrementally sample by sample in the presence of several types of changes and simultaneously retain old information in recurring scenarios. The key idea is to keep a moving window that slides when a new sample is available. The error of each model on the current window is determined using a boosting strategy that assigns small errors to the models that predict accurately the samples predicted poorly by the ensemble. To handle recurring and non-recurring changes, OWE uses a new assignment of models' weights that takes into account the models' errors on the past and current windows using a discounting factor that decreases or increases the contribution of old windows. In addition, OWE launches new models if the system's accuracy is decreasing, and it can exclude inaccurate models over time. Experiments with artificial and industrial data reveal that in most cases OWE outperforms other state-of-the-art concept drift approaches.

show abstract

Comparison of a genetic algorithm and simulated annealing for automatic neural network ensemble development

Cited by 50 publications

References 44 publications

Can multilayer perceptron ensembles model the ecological niche of freshwater fish species?

Can multilayer perceptron ensembles model the ecological niche of freshwater fish species?

Optimization of Multiple Traveling Salesman Problem Based on Simulated Annealing Genetic Algorithm

An on-line weighted ensemble of regressor models to handle concept drifts

Contact Info

Product

Resources

About