Learning kernel logistic regression in the presence of class label noise
作者：
Highlights：
• We propose an algorithm to learn the new robust multiple kernel logistic regression.
• The algorithm bypasses cross validation which is sub-optimal in noisy label settings.
• The algorithm is significantly faster than traditional cross validation approach.
• We empirically show that symmetric label noise can be as harmful as asymmetric noise.
摘要
Highlights•We propose an algorithm to learn the new robust multiple kernel logistic regression.•The algorithm bypasses cross validation which is sub-optimal in noisy label settings.•The algorithm is significantly faster than traditional cross validation approach.•We empirically show that symmetric label noise can be as harmful as asymmetric noise.
论文关键词：LR,Logistic Regression,rLR,robust Logistic Regression,KLR,Kernel Logistic Regression,rKLR,robust Kernel Logistic Regression,rMKLR,robust Multiple Kernel Logistic Regression,Classification,Label noise,Model selection,Multiple Kernel Learning
论文评审过程：Received 20 October 2012, Revised 8 May 2014, Accepted 11 May 2014, Available online 21 May 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.05.007
