摘要
根据主动学习可以有效地减少标注样本的代价这一特点,设计了一种基于SVM的主动学习方法.仿真实验中,检验分类正确率和F测度这两类评估指标,结果表明基于SVM的主动学习的学习效果优于被动学习.
As the active learning can reduce the cost of sample labeling effectively, we design an active learning method which is based on SVM. The simulation experiments show that the results of active learning method are much better than those of passive learning method not ouly in classification accuracy but also in F - Score.
出处
《长沙大学学报》
2014年第2期35-38,共4页
Journal of Changsha University
基金
国家自然科学基金(批准号:40972205)资助项目