International journal of advanced smart convergence

The Institute of Internet, Broadcasting and Communication (한국인터넷방송통신학회)
권호 표지
DOI QR코드

DOI QR Code

Model Selection in Artificial Neural Network

  • Received : 2018.09.19
  • Accepted : 2018.10.05
  • Published : 2018.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Artificial neural network is inspired by the biological neural network. For simplicity, in computer science, it is represented as a set of layers. Many research has been made in evaluating the number of neurons in the hidden layer but still, none was accurate. Several methods are used until now which do not provide the exact formula for calculating the number of thehidden layer as well as the number of neurons in each hidden layer. In this paper model selection approach was presented. Proposed model is based on geographical analysis of decision boundary. Proposed model selection method is useful when we know the distribution of the training data set. To evaluate the performance of the proposed method we compare it to the traditional architecture on IRIS classification problem. According to the experimental result on Iris data proposed method is turned out to be a powerful one.

Keywords

OTNBCL_2018_v7n4_57_f0001.png 이미지

Figure 1. XOR problem

OTNBCL_2018_v7n4_57_f0002.png 이미지

Figure 2. Nonlinear classifier is necessary in XOR problem

OTNBCL_2018_v7n4_57_f0003.png 이미지

Figure 3. Each perceptron produces a line in XOR problem

OTNBCL_2018_v7n4_57_f0004.png 이미지

Figure 4. Classifier architecture proposed by guidelines

OTNBCL_2018_v7n4_57_f0005.png 이미지

Figure 5. More complex problem

OTNBCL_2018_v7n4_57_f0006.png 이미지

Figure 6. Classification boundary in more complex problem

OTNBCL_2018_v7n4_57_f0007.png 이미지

Figure 7. Four lines and 2 in-between lines are necessary to classify

OTNBCL_2018_v7n4_57_f0008.png 이미지

Figure 8. Generated network architecture

OTNBCL_2018_v7n4_57_f0009.png 이미지

Figure 9. Distribution of IRIS data set according to each attribute

Table 1. Experimential results on traditional method and proposed one

OTNBCL_2018_v7n4_57_t0001.png 이미지

References

  1. S. Asthana, R.K. Bhujade, "Handwritten Multiscript Pin Code Recognition System having Multiple hidden layers using Back Propagation Neural Network", Journal of Electronics Communication and Computer Engineering, Vol. 2, No. 1 2011
  2. http://www.heatonresearch.com/node/707
  3. Y. Liu, J. A. Starzyk, Z. Zhu, "Optimizing Number Of Hidden Neurons in Neural Networks", http://www.ohio.edu/people/starzykj/network/Research/Papers/Recent%20conferences../Hidden%20Neurons%20AIA2007_549-204.pdf
  4. I.Rivals, L.Personnaz "A statistical procedure for determining the optimal number of hidden neurons of a neuralmodel", Second International Symposium on Neural Computation (NC'2000), Berlin, May 23-26, 2000.
  5. F.Fnaiech, N.Fnaiech, M.Najim, "A new feedforward neural network hidden layer neuron pruning algorithm" IEEE International Conference on Acoustics, Speech, and Signal Processing, 2001.
  6. K. Shinike, "A Two Phase Method for Determining the Number of Neurons in the Hidden Layer of a 3-Layer Neural Network", SICE Annual Conference 2010, August 18-21, 2010, The Grand Hotel, Taipei, Taiwan.
  7. https://www.linkedin.com/pulse/beginners-ask-how-many-hidden-layersneurons-use-artificial-ahmed-gad?fbclid=IwAR2p Kz8hTj82n3XPt5v2LrVStMF2VZJPQRSjVlb_RfM1AZIup5es1mnWhc8
  8. https://archive.ics.uci.edu/ml/datasets/iris
  9. S.B. Lee, H.G. Kim, H.K.Seok, J.H. Nang, "Comparison of Fine-Tuned Convolutional Neural Networks for Clipart Style Classification" International Journal of Internet, Broadcasting and Communication(IJIBC), Vol.9 No.4, pp.1-7, 2017 DOI: https://doi.org/10.7236/IJIIBC.2017.9.4.1