论文部分内容阅读
基于不确定性最大的主动学习是一种普遍应用的主动学习算法,该方法选择当前分类器最难确定类别的样例。本文提出了一种基于样例池不确定性缩减最大的主动学习方法,该方法选择那些能够使得样例池不确定性缩减最大的样例,从而使得到的分类器具有更好的泛化能力。
Based on the most uncertain active learning is a commonly used active learning algorithm, the method to select the most difficult to determine the class of the current classifier examples. In this paper, an active learning method based on the largest reduction of the uncertainty of the sample pool is proposed. The method chooses samples that can minimize the uncertainty of the sample pool, so that the resulting classifier has better generalization ability .