改进的集成神经网络学习优化研究

Research on the optimization of neural network ensemble learning

AdaBoosting算法是一种集成学习算法,用来提高不稳定学习器的泛化能力.本文基于神经网络是一种不稳定分类器的机理结合最小二乘算法和遗传算法,提出了对集成神经网络分类器权值进行优化的新方法.

Ensemble learning is an algorithm to improve the generalization ability of the unstable classifier. Based on the unstable classifier of NN and combined the Least Square Algorithm（LSA） and Genetic Algorithm（GA）, the new method has been proposed to optimize the weights of ensemble neural network classifiers.