Online Learning of Approximate Maximum Margin Classifiers with Biases
Online Learning of Approximate Maximum Margin Classifiers with Biases
复制标题
DOI：
--
复制
发表时间：
2007
期刊：
影响因子：
--
通讯作者：
Kosuke Ishibashi-;Kohei Hatano-;Masayuki Takeda
中科院分区：
文献类型：
--
作者：
Kosuke Ishibashi-;Kohei Hatano-;Masayuki Takeda
我们考虑线性分类器的在线学习，近似最大化2-范数的利润。给定一个线性可分离的实例序列，典型的在线学习算法（如Perceptron及其变体）将它们映射到具有额外维度的增强空间中，以便这些实例被线性分类器分离，而没有恒定的偏置项。但是，此映射可能会减少实例上的余量。在本文中，我们提出了一个修改后的版本，李和龙的ROMMA，避免这样的映射，我们表明，我们的修改后的算法实现了更高的利润率比以前的在线学习算法。
We consider online learning of linear classifiers which approximately maximize the 2-norm margin. Given a linearly separable sequence of instances, typical online learning algorithms such as Perceptron and its variants, map them into an augmented space with an extra dimension, so that those instances are separated by a linear classifier without a constant bias term. However, this mapping might decrease the margin over the instances. In this paper, we propose a modified version of Li and Long’s ROMMA that avoids such the mapping and we show that our modified algorithm achieves higher margin than previous online learning algorithms.