欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
给出了Tikhonov正则算法的多核正则分类器在α混合样本下的错分误差的界.利用α混合样本的Bernstein不等式,得到了α混合样本情况下的两个相关结果,然后将这两个结果应用到误差分解,并得到学习速率的估计.
英文摘要：
Tikhonov regularization scheme is one of the most common classification algorithms.Our main purpose was to estimate the bound of excess misclassification error of multi-kernel classfiers with α mixing observations based on the algorithms.Using Bernstein inequality for α mixing sequences,two related results with α mixing observarions were obtained.Furthermore,the two results were applied to error decomposition and the learning rates are implied.
同期刊论文项目
同项目期刊论文
期刊信息
位置：