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：
根据李群具有微分流形的性质,提出了一种李群机器学习的线性分类算法.该分类方法的思想是：首先将样本数据集嵌入到微分流形当中,每个实例对应着流形上的一个点,利用测地线距离度量两个实例间的距离;然后,根据测试实例和训练数据集中实例间测地线距离的代数关系,确定测试实例的分类.实验测试表明,该线性分类算法较k-最近邻算法（KNN）及NaiveBayes分类算法具有较高的分类精度.
英文摘要：
According to Lie group has shaped the nature of differential manifold, a linear Lie group machine learning classification algorithms is proposed. The idea of the classification is： first of all, sample data sets will be embedded into the differential manifold. Each instance is corresponding to point on the manifold. Make use of geodesic distance to measure the distance between two examples; and then in accordance with the algebraic relationship between test examples and training data, design a classifier to determine the classification of test examples. Experimental tests show that the Lie group of linear classification machine learning algorithm has higher classification accuracy.
同期刊论文项目
同项目期刊论文
期刊信息
位置：