基于支持向量机与概率输出网的深度学习模型

本文提出了基于支持向量机与概率输出网络的深度学习回归模型.该回归模型利用深度学习的深层结构,以及支持向量机的泛化能力、概率输出网络中的条件概率估计特点,建立了多层支持向量机的深度学习结构,避免了深度学习的参数选择问题.其中核参数的选择域呈网格状,通过求取输出对应β分布的累积概率分布和经验累积概率分布的K-S检验,求取一致性的p值最大对应的核参数作为支持向量机模型的核参数.对应的输出为模型提取的特征,作为下一层的输入,直至模型达到结束条件为止.仿真实验通过三个标准的回归数据集证明了本文提出模型的有效性.

Deep learning model of support vector machines with probability output networks

In this paper, a deep learning regression model based on support vector machines and probabilistic output networks is proposed. Based on the deep structure of deep learning, generalization ability of support vector machines (SVM) and conditional probability estimation in probability output network, a multi-layer SVM is established. The problem of parameter selection in deep learning is avoided. Where the kernel widths were constructed as the form of a grid. The kernel parameters are obtained according to the maximum value of the p-values where chosen by the K-S test of the cumulative probability distribution and the empirical cumulative probability distribution of the corresponding β distribution. The corresponding output is the extracted feature, which is used as the input of the next layer model until the model reaches the end condition. The simulation experiments prove the effective of the proposed model by three standard regression data set.