A new multilayer feedforward network based on trimmed mean neuron model

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Published

Standard

A new multilayer feedforward network based on trimmed mean neuron model. / Yolcu, Ufuk; Bas, Eren; Egrioglu, Erol et al.
In: Neural Network World, Vol. 25, No. 6, 01.01.2015, p. 587-602.

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Harvard

Yolcu, U, Bas, E, Egrioglu, E & Aladag, CH 2015, 'A new multilayer feedforward network based on trimmed mean neuron model', Neural Network World, vol. 25, no. 6, pp. 587-602. https://doi.org/10.14311/NNW.2015.25.029

APA

Yolcu, U., Bas, E., Egrioglu, E., & Aladag, C. H. (2015). A new multilayer feedforward network based on trimmed mean neuron model. Neural Network World, 25(6), 587-602. https://doi.org/10.14311/NNW.2015.25.029

Vancouver

Yolcu U, Bas E, Egrioglu E, Aladag CH. A new multilayer feedforward network based on trimmed mean neuron model. Neural Network World. 2015 Jan 1;25(6):587-602. doi: 10.14311/NNW.2015.25.029

Author

Yolcu, Ufuk ; Bas, Eren ; Egrioglu, Erol et al. / A new multilayer feedforward network based on trimmed mean neuron model. In: Neural Network World. 2015 ; Vol. 25, No. 6. pp. 587-602.

Bibtex

@article{d32de81f4e6942d3a0e05b8c73fb2c6c,

title = "A new multilayer feedforward network based on trimmed mean neuron model",

abstract = "The multilayer perceptron model has been suggested as an alternative to conventional approaches, and can accurately forecast time series. Additionally, several novel artificial neural network models have been proposed as alternatives to the multilayer perceptron model, which have used (for example) the generalized- mean, geometric mean, and multiplicative neuron models. Although all of these artificial neural network models can produce successful forecasts, their aggregation functions mean that they are negatively affected by outliers. In this study, we propose a new multilayer, feed forward neural network model, which is a robust model that uses the trimmed mean neuron model. Its aggregation function does not depend on outliers. We trained this multilayer, feed forward neural network using modified particle swarm optimization. We applied the proposed method to three well-known time series, and our results suggest that it produces superior forecasts when compared with similar methods.",

keywords = "Forecast, Neural networks, Neuron model, Outliers, Particle swarm optimization, Trimmed mean",

author = "Ufuk Yolcu and Eren Bas and Erol Egrioglu and Aladag, {Cagdas Hakan}",

year = "2015",

month = jan,

day = "1",

doi = "10.14311/NNW.2015.25.029",

language = "English",

volume = "25",

pages = "587--602",

journal = "Neural Network World",

issn = "1210-0552",

publisher = "Czech Technical University in Prague",

number = "6",

}

RIS

TY - JOUR

T1 - A new multilayer feedforward network based on trimmed mean neuron model

AU - Yolcu, Ufuk

AU - Bas, Eren

AU - Egrioglu, Erol

AU - Aladag, Cagdas Hakan

PY - 2015/1/1

Y1 - 2015/1/1

N2 - The multilayer perceptron model has been suggested as an alternative to conventional approaches, and can accurately forecast time series. Additionally, several novel artificial neural network models have been proposed as alternatives to the multilayer perceptron model, which have used (for example) the generalized- mean, geometric mean, and multiplicative neuron models. Although all of these artificial neural network models can produce successful forecasts, their aggregation functions mean that they are negatively affected by outliers. In this study, we propose a new multilayer, feed forward neural network model, which is a robust model that uses the trimmed mean neuron model. Its aggregation function does not depend on outliers. We trained this multilayer, feed forward neural network using modified particle swarm optimization. We applied the proposed method to three well-known time series, and our results suggest that it produces superior forecasts when compared with similar methods.

AB - The multilayer perceptron model has been suggested as an alternative to conventional approaches, and can accurately forecast time series. Additionally, several novel artificial neural network models have been proposed as alternatives to the multilayer perceptron model, which have used (for example) the generalized- mean, geometric mean, and multiplicative neuron models. Although all of these artificial neural network models can produce successful forecasts, their aggregation functions mean that they are negatively affected by outliers. In this study, we propose a new multilayer, feed forward neural network model, which is a robust model that uses the trimmed mean neuron model. Its aggregation function does not depend on outliers. We trained this multilayer, feed forward neural network using modified particle swarm optimization. We applied the proposed method to three well-known time series, and our results suggest that it produces superior forecasts when compared with similar methods.

KW - Forecast

KW - Neural networks

KW - Neuron model

KW - Outliers

KW - Particle swarm optimization

KW - Trimmed mean

U2 - 10.14311/NNW.2015.25.029

DO - 10.14311/NNW.2015.25.029

M3 - Journal article

AN - SCOPUS:84987760533

VL - 25

SP - 587

EP - 602

JO - Neural Network World

JF - Neural Network World

SN - 1210-0552

IS - 6

ER -

Research

Links

Text available via DOI:

Keywords