一种基于AdaBoost的极限学习机分类方法

极限学习机是一种新型的单隐层前馈神经网络,在训练网络的过程中随机给定输入层权值和隐藏层偏差,所以训练速度非常快,但却导致了输出不稳定.提出了一种基于AdaBoost的极限学习机,把极限学习机作为AdaBoost的基本分类器,通过改变输入数据的权重,使得极限学习机的分类性能得到提升.实验结果表明了该方法与极限学习机和传统的神经网络相比,能够提高极限学习机的学习性能,并且使极限学习机输出更加稳定.

A Classification Method of Extreme Learning Machine Based on AdaBoost

Extreme learning machine was a new single hidden layer feedforward neural network.In the training process of the network,input layer and hidden layer deviation were given randomly.So the training speed was very fast.But because of it,the output of extreme learning machine was unstable.An AdaBoost-based extreme learning machine was presented.By changing the weights of input data,the performance of extreme learning machine could be improved.Experimental results showed that the proposedalgorithm achieved better performance and acted more stably than similar methods.