Third International conference on advances in communication, network and computing

Research Article

Multilayer Feed-Forward Artificial Neural Network Integrated with Sensitivity Based Connection Pruning Method

Download
248 downloads
  • @INPROCEEDINGS{10.1007/978-3-642-35615-5_10,
        author={Siddhaling Urolagin and Prema K.V. and JayaKrishna R. and N. Reddy},
        title={Multilayer Feed-Forward Artificial Neural Network Integrated with Sensitivity Based Connection Pruning Method},
        proceedings={Third International conference on advances in communication, network and computing},
        proceedings_a={CNC},
        year={2012},
        month={12},
        keywords={Sensitivity measure pruning algorithm generalization global error},
        doi={10.1007/978-3-642-35615-5_10}
    }
    
  • Siddhaling Urolagin
    Prema K.V.
    JayaKrishna R.
    N. Reddy
    Year: 2012
    Multilayer Feed-Forward Artificial Neural Network Integrated with Sensitivity Based Connection Pruning Method
    CNC
    Springer
    DOI: 10.1007/978-3-642-35615-5_10
Siddhaling Urolagin1,*, Prema K.V.2,*, JayaKrishna R.1,*, N. Reddy2,*
  • 1: M.I.T.
  • 2: Mody Institute of Technology and Science
*Contact email: siddesh_u@yahoo.com, prema_kv@rediffmail.com, jayakrishnaa.r@gmail.com, dr_nvsreddy@rediffmail.com

Abstract

The Artificial Neural Network (ANN) with small size may not solve the problem while the network with large size will suffer from poor generalization. The pruning methods are approaches for finding appropriate size of the network by eliminating few parameters from the network. The sensitivity based pruning will determine sensitivity of the network error for removal of a parameter and eliminate parameters with least sensitivity. In this research a sensitivity based pruning method is integrated with multilayer feed-forward ANN and applied on MNIST handwritten numeral recognition. An analysis of effect of pruning on the network is compared with performance of a network without pruning. It is observed that the network integrated with pruning method show better generalization ability than a network without pruning method being incorporated.