DocumentCode :
2337847
Title :
Analysis of sensitivity behavior of Madalines
Author :
Wang, Ying-Feng ; Zeng, Xiao-qin ; Yeung, Daniel S.
Author_Institution :
Dept. of Comput. Sci. & Eng., Hohai Univ., Nanjing, China
Volume :
8
fYear :
2005
fDate :
18-21 Aug. 2005
Firstpage :
4731
Abstract :
This paper aims at exploring the behavior of the sensitivity to weight perturbation for an ensemble of Madalines. An analytical formula is derived first for the calculation of the Adalines´ sensitivity, and then based on it an algorithm is given for the computation of the Madalines´ sensitivity. The efficiency of the computation is verified by computer simulations. By means of the formula and the algorithm, the sensitivity tendency is analyzed with regard to the variation of some parameters, such as the dimension of input, the number of Adalines in a layer and the number of layers. The analysis results show that the dimension of input has little effect on the sensitivity as long as the dimension is sufficient large, while the increases in the number of Adalines in a layer and the number of layers will lead the sensitivity to increase under an upper bound. These tendencies of the Madalines´ sensitivity will be useful for designing robust Madalines with appropriate architecture.
Keywords :
learning (artificial intelligence); neural nets; Adalines; Madalines; artificial neural network; computer simulation; sensitivity behavior analysis; weight perturbation; Algorithm design and analysis; Artificial neural networks; Computer science; Computer simulation; Hypercubes; Neural networks; Probability; Robustness; Sensitivity analysis; Upper bound; Adaline; Madaline; neural networks; sensitivity analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
Type :
conf
DOI :
10.1109/ICMLC.2005.1527774
Filename :
1527774
Link To Document :
بازگشت