局部感知的类限制极限学习机

极限学习机(ELM)由于高效的训练方式被广泛应用于分类回归,然而不同的输入权值在很大程度上会影响其学习性能。为了进一步提高ELM的学习性能,针对ELM的输入权值进行了研究,充分利用图像局部感知的稀疏性,将局部感知的方法运用到基于自动编码器的ELM(ELM-AE)上,提出了局部感知的类限制极限学习机(RF-C₂ELM)。通过对MNIST数据集进行分类问题分析实验,实验结果表明,在具有相同隐层结点数的条件下,提出的方法能够获得更高的分类精度。

Receptive field class-constrained extreme learning machine

Because of efficient training methods, the extreme learning machine (ELM) was widely used in classification and regression. However different input weights largely affected the learning performance. To further improve the learning performance of ELM, the paper applied local receptive method to ELM based on auto encoder (ELM-AE), and proposed the algorithm called receptive field class-constrained extreme learning machine (RF-C₂ELM). Based on the analysis of the classification problem of MNIST data set, the experimental results show that the proposed method can obtain higher classification accuracy under the same number of hidden nodes.