基于神经网络集成的数据分析

神经网络集成通过训练多个神经网络并将其结论进行适当的合成,可以显著地提高学习系统的泛化能力.然而,设计一个好的神经网络集成必须在个体准确性与彼此差异性之间取得一个平衡.本文提出了一种改进的神经网络集成构造方法--基于噪声传播的神经网络集成算法(NSENN).

Data Analysis based on Ensemble of Neural Networks

Neural network ensemble is a learning paradigm where many neural networks are jointly used to solve a problem. It can significantly improve the generalization ability of learning systems. For an effective neural network ensemble, it is critical that the ensemble be made up of relatively accurate individual networks that are very different from each other. Thus, we must make a trade-off between accuracy and diversity. In this paper, we propose an improved constructing approach, i.e. noisy spread based ensemble of neural networks (NSENN).