论文部分内容阅读
针对选择性集成逆向传播神经网络(GASEN-BPNN)模型训练学习速度慢,选择性集成极限学习机(GASEN-ELM)模型建模精度稳定性差等问题,提出一种基于遗传算法的选择性集成核极限学习机(GASEN-KELM)建模方法。该方法首先通过对训练样本进行随机采样获取子模型训练样本;然后采用泛化性、稳定性较佳的核极限学习机(KELM)算法建立候选子模型,通过标准遗传算法工具箱,依据设定阈值按进化策略优化选择最佳子模型;最后通过简单平均加权集成的方式获得最终GASEN-KELM模型。采用标准混凝土抗压强度数据验证了所提出方法的有效性,并与GASEN-BPNN和GASEN-ELM选择性集成算法进行比较,表明所提出方法可以在模型学习速度和建模预测稳定性方面获得较好的均衡。
In view of the problems of slow learning and training of GASEN-BPNN model and poor stability of GASEN-ELM model, this paper proposes a genetic algorithm based on selective integration kernel Limit learning machine (GASEN-KELM) modeling method. Firstly, the training samples are obtained by random sampling to obtain the sub-model training samples. Then, the candidate sub-models are established by the kernel-based learning machine (KELM) with generalization and stability. Through the standard genetic algorithm toolbox, Finally, the final GASEN-KELM model is obtained by simple average weighted integration. Comparing with the GASEN-BPNN and GASEN-ELM selective integration algorithms, the proposed method is validated by the standard concrete compressive strength data. The results show that the proposed method can obtain better performance in terms of model learning speed and modeling predictive stability Well balanced.