Title :
A comparison of sensitivity definitions based on expectation and variance
Author :
Xu, Ning ; Li, Chun-guo
Author_Institution :
Sch.´´s Conciliation Board, Hebei Univ., Baoding, China
Abstract :
The sensitivity analysis plays a very important role in the construct of a network and its application. There has been two statistical definitions based on expectation and variance respectively. This paper compares the two definitions and tries to find which definition is more suitable to define the sensitivity of one network. According to the immediate aspect and simulation experiments, mathematical expectation fits the definition more accurately.
Keywords :
mathematical analysis; sensitivity analysis; statistical analysis; mathematical expectation; sensitivity analysis; sensitivity definition; statistical definition; Biological neural networks; Cybernetics; Machine learning; Probability distribution; Random variables; Sensitivity analysis; MLP; Mathematical expectation; Sensitivity definition; variance;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2011 International Conference on
Conference_Location :
Guilin
Print_ISBN :
978-1-4577-0305-8
DOI :
10.1109/ICMLC.2011.6016814
