hessian正则化logistic回归模型-计算机工程与应用.pdfVIP

236 2016 ，52（5 ） Computer Engineering and Applications 计算机工程与应用 Hessian 正则化Logistic 回归模型 刘红丽，刘伟锋，王延江，董丽萍 LIU Hongli, LIU Weifeng, WANG Yanjiang, DONG Liping 中国石油大学 信息与控制工程学院，山东 青岛 266580 College of Information and Control Engineering, China University of Petroleum, Qingdao, Shangdong 266580, China LIU Hongli, LIU Weifeng, WANG Yanjiang, et al. Hessian regularized Logistic regression. Computer Engineering and Applications, 2016, 52 （5 ）：236-240. Abstract ：The Semi-Supervised Learning （SSL）method of Laplacian regularization opens a bright avenue by studying the classification problem of a large number of images. However, Laplacian method biases the classification function toward a constant function and possibly results in poor generalization. So the Hessian regularized Logistic Regression （HesLR ） method for image classification is proposed. Hessian regularization can well predict the data points beyond the boundary of the domain. Extensive experiments on the MIR Flickr dataset validate the effectiveness of the proposed method by com- paring it with baseline algorithms, including SVM, the Logistic regression method and the Laplacian regularized Logistic regression method. Key words ：Semi-Supervised Learning（SSL）; Hessian; kernel Logistic regression; image classification 摘 要：针对大数据量的图像分类问题，Laplacian 正则化的半监督学习方法获得了广阔的应用前景。然而Laplacian 正则化使分类函数趋向于常数函数而易导致较差的推测能力。提出了基于Hessian 正则化的Logistic 回归模型用于 图像分类，Hessian 正则化可以较好地预测区域之外的数据点。在MIR Flickr 数据库上进行图像分类实验，与SVM 、 Logistic 回归和Laplacian 正则化的Logistic 回归方法相比，Hessian 正则化的Logistic 回归模型更有效。 关键词：半监督学习；Hessian ；Logistic 核回归；图像分类 文献标志码：A 中图分类号：TP391.41 doi ：10.3778/j.issn. 1002-8331.1402-0363 1 引言 入算法（LLE ），使降维后的数据保持原有的拓扑结构。 随着数码相机、电脑硬件、互联网以及机器学习等技 Josh Tenenbaum 在文[3] 中提出等距映射（Isomap ）方法， 术的不断发展，百万甚至千万的图像的获取及处理成为 采用测地线距离计算高维流形数据点间距。Donoho