University of Guilan
University of Guilan Press

 Statistics

Number of Journals32
Number of Issues840
Number of Articles8,153
Article View52,502,735
PDF Download8,896,678
Sobhani, J., sobhani, Z., Sheikhan, M. (2008). Neural networks for predicting the compressive strength of concrete: error back-propagation and recurrent elman networks. Aquatic Physiology and Biotechnology, 1(2), 19-33.
Jafar Sobhani; zahra sobhani; Mansour Sheikhan. "Neural networks for predicting the compressive strength of concrete: error back-propagation and recurrent elman networks". Aquatic Physiology and Biotechnology, 1, 2, 2008, 19-33.
Sobhani, J., sobhani, Z., Sheikhan, M. (2008). 'Neural networks for predicting the compressive strength of concrete: error back-propagation and recurrent elman networks', Aquatic Physiology and Biotechnology, 1(2), pp. 19-33.
Sobhani, J., sobhani, Z., Sheikhan, M. Neural networks for predicting the compressive strength of concrete: error back-propagation and recurrent elman networks. Aquatic Physiology and Biotechnology, 2008; 1(2): 19-33.

Neural networks for predicting the compressive strength of concrete: error back-propagation and recurrent elman networks

تحقیقات بتن
Article 2, Volume 1, Issue 2, July 2008, Pages 19-33    PDF (598.61 K)
Authors
Jafar Sobhani* ; zahra sobhani; Mansour Sheikhan
Abstract
In the recent years, the artificial neural networks (NNT) have been widely applied in the various fields of engineering especially in the various fields of the civil engineering. In this paper, two types of neural networks with three architectures were used to predict the compressive strength of concrete samples. In this study, a novel type of NNT, named as the recurrent Elman networks, was introduced and used to predict the compressive strength of concrete samples. Moreover, in this paper, the results of simulation with the Elman networks were compared with the results of traditional back propagation networks. The results of comparison showed that the two layered Elman network which has 5 and 3 neurons in the first and second layer respectively, has the best performance from the generalization perspective; and vice versa the standard BP (with 8 and 5 neurons in the first and second layer) has got the best performance for the estimation purposes
Statistics
Article View: 4,677
PDF Download: 2,985


Journal Management System. Powered by Sinaweb