基于样本重构的神经网络集成学习方法

在回顾以往神经网络集成的研究成果基础上，提出一种新的负相关学习方法，该方法易于执行，计算量小，有效的消除了学习中的复合线性问题，减小了集成误差，最后用测试用例对该方法进行了考察，证明该方法可以有效的降低集成预测误差，得到较为理想的集成效果。

Learning approach for neural network ensemble based on specimen reconstruct

Based the review of the research of neural network ensemble,a new negative correlation learning method was proposed,which was both easy to implement and had the less computation,the co-linearity problem was eliminated,the error of the neural network ensemble was reduced.In the end,gives an test to this method,which prove that the method can reduce the error of the neural network ensemble effectively and get an expected result.