改进的加权极速学习机

针对加权极速学习机人为固定权重可能会错失更优权重的问题,提出了改进的加权极速学习机。该方法的多数类的初始权重设为1,使用多数类与少数类样例数的比值作为少数类的初始权重,然后通过在多数类或者少数类中添加权重调节因子,从缩小和扩大两个方向去调节权重,最后通过实验结果选出最优的权重。实验分别使用原加权极速学习机、其他权重的极速学习机和新方法在改造的UCI数据集上进行比较。结果表明新方法无论是在F-mea-sure还是G-mean上都要优于其他加权极速学习机。

Improved Weighted Extreme Learning Machine

If the weight of the original weighted extreme learning machine is fixed artificially,the more optimal weight may be missed.Aiming at this problem,an improved weighted extreme learning machine was proposed.The new method uses the ratio of the sample number of different classes as the initial weight.The weight is adjusted by the weight adjustment factor from two directions of reducing and enlarging the weight ratio.Finally,the optimal weights are selected by the experimental results.The experiments were carried out on the transformed UCI data set using the original weighted extreme learning machine,the weighted extreme learning machine with other weights and the new method respectively.The experimental results indicate that the improved weighted extreme learning machine has better classification performance.