说明：  极限学习机( ELM) 在训练过程中无须调整隐层节点参数，因其高效的训练方式被广泛应用于分类和回 归，然而极限学习机也面临着结构选择与过拟合等严重问题。为了解决此问题，针对隐层节点增量数目对收敛 速度以及训练时间的影响进行了研究，提出一种利用网络输出误差的变化率控制网络增长速度的变长增量型极 限学习机算法( VI-ELM) 。通过对多个数据集进行回归和分类问题分析实验，结果表明，提出的方法能够以更高 效的训练方式获得良好的泛化性能。
(The extreme learning machine (ELM) does not need to adjust the hidden node parameters during the training process, because its efficient training method is widely used in classification and regression. However, the extreme learning machine also faces serious problems such as structure selection and overfitting. In order to solve this problem, the influence of the number of hidden layer nodes on the convergence speed and training time is studied, and a variable length incremental limit learning machine algorithm that uses the rate of change of the network output error to control the network growth rate is proposed.)