Google scholar arxiv informatics ads IJAIS publications are indexed with Google Scholar, NASA ADS, Informatics et. al.

Call for Paper

-

May Edition 2020

International Journal of Applied Information Systems solicits high quality original research papers for the May 2020 Edition of the journal. The last date of research paper submission is April 15, 2020.

An Improved Approach for Hidden Nodes Selection in Artificial Neural Network

H. N. Odikwa, Nkechi Ifeanyi-Reuben, Osaki Miller Thom-Manuel in Artificial Intelligence

International Journal of Applied Information Systems
Year of Publication: 2020
Publisher: Foundation of Computer Science (FCS), NY, USA
Authors:H. N. Odikwa, Nkechi Ifeanyi-Reuben, Osaki Miller Thom-Manuel
10.5120/ijais2020451837
Download full text
  1. H N Odikwa, Nkechi Ifeanyi-Reuben and Osaki Miller Thom-Manuel. An Improved Approach for Hidden Nodes Selection in Artificial Neural Network. International Journal of Applied Information Systems 12(27):7-14, February 2020. URL, DOI BibTeX

    @article{10.5120/ijais2020451837,
    	author = "H. N. Odikwa and Nkechi Ifeanyi-Reuben and Osaki Miller Thom-Manuel",
    	title = "An Improved Approach for Hidden Nodes Selection in Artificial Neural Network",
    	journal = "International Journal of Applied Information Systems",
    	issue_date = "February 2020",
    	volume = 12,
    	number = 27,
    	month = "February",
    	year = 2020,
    	issn = "2249-0868",
    	pages = "7-14",
    	url = "http://www.ijais.org/archives/volume12/number27/1077-2020451837",
    	doi = "10.5120/ijais2020451837",
    	publisher = "Foundation of Computer Science (FCS), NY, USA",
    	address = "New York, USA"
    }
    

Abstract

Training of data in Artificial Neural Network (ANN) imposes a lot of difficulty when determining the number of hidden layers and nodes that will enhance the network convergence in a multi layer perceptron neural network. This study employed a different approach of choosing hidden layers and nodes by taken cognizance of the fact that it is the bedrock of easy network convergence in an artificial neural network by using a heuristic search function of means end analysis. The system adopted the means end analysis algorithm by using a forward and backward chaining to generate current operators and calculating their differences from the goal state which is the target value of the ANN. The system was trained using 500 prostate data and 100 diabetes patient diseases from Federal Medical Center Umuahia in Abia State, Nigeria to train the data in a neural network. The trained data was used for classification. The result revealed that means end analysis is promising to training data in an ANN and yielded accuracies 80%, 82%, 85% for hidden layers between 2 to 20 and hidden nodes between 2 to 6. The classification accuracies of 87%, 90%, 95%, 98% for prostate cancer disease were obtained for hidden layers of 30 to 60 and hidden nodes between 8 to 14. The classification accuracies for diabetes disease were 81%, 85%, 87% for hidden layers of 30 to 60 and 89%, 92.5%, 95%, 98.5% for hidden nodes between 8 to 14 for diabetes disease considering time and space trade-offs.

Reference

  1. Adel, E. (2009). Advanced applications for artificial neural network. Research Gate. Intech
  2. Alan, J.T., Miltos, P., Simon, D.W., Saeed, M.G & Robert, E.M. (2018). Two hidden layers are usually better than one. Institute of Electrical and Electronics Engineering. 2(4): 244-252.
  3. Benson-Emenike, M. E. & Ifeanyi-Reuben, N. J. (2018). Network Learning and Training of a Cascaded Link-Based Feed Forward Neural Network (CLBFFNN) in an Intelligent Trimodal Biometric System. International Journal of Artificial Intelligence and Applications (IJAIA). 9(6): 27 – 47
  4. David, R. (2005). Neural network architecture development process. International journal of Adhoc, sensors and ubiquitous computing. 10(7): 123-128.
  5. Fabian, O., Kachi, G., & Okoli, H, B. (2010). Prospects of feed forward algorithm in Multilayer layer perceptron. International journal of grid computing and applications. 12(10): 34 – 40.
  6. Hebtamu, Z.A., Wei, W. & Junhong, Z. (2018). Feed-forward Neural Networks with a Hidden layer regularization Method. Journal of MDPI. 4(12).
  7. Imran, S., Ahmad, J., Syed, I.S & Faisal, M.K. (2016). Impact of Varying nodes and hidden layers in neural network architecture for a time frequency application. Conference multi-topic of Institute of Electrical and Electronics engineering.
  8. Odikwa, H., Ugwu, C. & Inyiama, H. (2017). An Improved Model for Clinical Decision Support System. International Journal of Artificial Intelligence and Applications (IJAIA). 8(6): 37-55.
  9. Panchal, F.S., & Panchal, M. (2014). Review on Methods of Selecting Number of Hidden Nodes in Artificial Neural Network. International journal of computer science and mobile computing. 3(11): 455-464.
  10. Saurabb, K.(2012). Approximating Number of Hidden layer Nodes in Multiple Hidden layer BPNN Architecture. International Journal of Engineering trends and technology. 3(6): 714-717
  11. Serena, H. C., Anthony, J. J. and John, P. N. (2008). Artificial Intelligence techniques: An introduction to their use for modelling environmental systems. Elsevier – Mathematics and Computers in Simulation. 78(2008): 379 - 400
  12. Sreekanth, R. K. (2012). Artificial Intelligence Algorithms. IOSR Journal of Computer Engineering (IOSRJCE. 6(3): 01-08.
  13. Stathakis, D.(2009). How many hidden layers and nodes. International journal of remote sensing. 30(8): 2133-2147.

Keywords

Artificial Neural Network, Multi-layer Perceptron, Artificial Intelligence, Hidden Layers, Hidden Nodes, Prostate Cancer, Diabetes