稀疏自编码和Softmax回归的快速高效特征学习

针对特征学习效果与时间平衡问题,提出了一种快速高效的特征学习方法.将稀疏自编码和Softmax回归组合成一个新的特征提取模型,在提取原始图像潜在信息的基础上,利用多分类器返回值可以反映输入信息的相似程度的特点,快速高效的学习利于分类的特征向量.鉴于标签信息已知,该算法在图像分类效果上明显优于几种典型的特征学习方法.为了使所提算法具有更好的泛化能力,回归模型的损失函数中加入了L2范数防止过拟合,同时,采用随机梯度下降的方法得到模型的最优参数.在4个标准数据集上的测试结果表明该算法是有效可行的.

Fast and efficient feature learning algorithm based on sparse autoencoder and Softmax regression

Aiming at equilibrium problem between the effectiveness and time of feature learning,a fast and efficient feature learning method is proposed.A new feature extraction model is combined with sparse autoencoder and sofimax regression.After extracting the potential information of original image,the model take advantage of return value of multiple classifier which can respond the similarity of input information,learning feature vector for classification fastly and efficiently.In view of the label information is known,this algorithm is better than several typical feature learning methods in the image classification.In order to make the proposed algorithm has better generalization ability,add L2-norm into loss function of regression model to prevent overfitting.At the same time,use stochastic gradient descent method to get the optimal parameters of the model The test results on four standard data sets show that the proposed algorithm is feasible and effective.