Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset

Abstract: Artificial neural network (ANN) is inspired from biological neural networks and has abilities like adaptability, ability to generalize and learning capability. An ANN can be used for a specific application such as pattern recognition, function approximation or data classification through a learning process. Back-propagation (BP) algorithm which is a gradient-descent method is the most common algorithm for training ANNs. However it has some disadvantages; it requires differentiable neuron transfer function and has a high possibility to converging into local optima. In order to overcome these shortcomings of BP algorithm meta-heuristic algorithms are successfully used for training ANNs in the literature. In this study, a new generation meta-heuristic namely Ions motion (IM) algorithm is used for training ANNs on a real-life real estate dataset. The dataset includes several attributes such as facade, number of rooms, number of bathrooms, building age, heating system, etc. The aim of the study is to estimate the sales price of a flat by taken into consideration these input variables. ANNs is trained by using IM algorithm on the real estate dataset and the error rates are obtained. The capability of the IM algorithm on training ANNs is compared with some regression-based algorithms and ANNs models from Weka data mining software. The obtained error rates and comparison with the other algorithms shows that the IM algorithm is highly capable on training ANNs on real estate dataset.

Keywords: Training artificial neural networks, Ions motion algorithm, Real estate dataset

Publication details: Conference abstract presented at the 3rd International Conference on Engineering and Natural Sciences (ICENS 2017), 3-7 May 2017, Budapest, Hungary. Published in: Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017), pp. 505. Publisher: ICENS. ISBN: 978-605-83575-4-9.

Acknowledgement: This study was supported by the Research Fund of Erciyes University (Project number: 6771).

How to cite / Kaynak gösterimi

APA 7: Gülmez, B., & Kulluk, S. (2017, May). Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset [Conference abstract]. In Ö. Çınar (Ed.), Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017) (p. 505). ICENS. 3rd International Conference on Engineering and Natural Sciences (ICENS 2017), Budapest, Hungary. https://doi.org/10.5281/zenodo.23036891

IEEE: B. Gülmez and S. Kulluk, "Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset," in Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017), Budapest, Hungary, May 2017, p. 505, ISBN: 978-605-83575-4-9, doi: 10.5281/zenodo.23036891.

Chicago: Gülmez, Burak, and Sinem Kulluk. 2017. "Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset." In Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017), 505. Budapest, Hungary: ICENS. https://doi.org/10.5281/zenodo.23036891.

MLA 9: Gülmez, Burak, and Sinem Kulluk. "Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset." Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017), ICENS, 2017, p. 505. https://doi.org/10.5281/zenodo.23036891.

Harvard: Gülmez, B. and Kulluk, S. (2017) 'Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset', in Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017). Budapest, Hungary: ICENS, pp. 505. doi: 10.5281/zenodo.23036891.

BibTeX:

@inproceedings{Gulmez2017training,
  author    = {Gülmez, Burak and Kulluk, Sinem},
  title     = {Training Artificial Neural Networks with Ions Motion Algorithm on Real Estate Dataset},
  booktitle = {Book of Abstracts of the III International Conference on Engineering and Natural Sciences (ICENS 2017)},
  year      = {2017},
  month     = {may},
  pages     = {505},
  address   = {Budapest, Hungary},
  publisher = {ICENS},
  isbn      = {978-605-83575-4-9},
  doi       = {10.5281/zenodo.23036891},
  url       = {https://doi.org/10.5281/zenodo.23036891}
}