摘要
典型的非参数估计方法有Parzen窗法和K近邻法两种。Parzen窗法的缺点是对训练样本的需求量较大,存在维数灾难问题,对此可通过并行神经网络结构实现并改进。K-近邻法一般遵循近邻规则——KNN,直接用来进行样本分类。
There are two typical non-parametric method,Parzen window method and K-nearest neighbor method.Disadvantages of Parzen window method are large demand for training samples,existing dimension disaster problems,which could improved by parallel neural network structure.K-nearest neighbor method generally follows the nearest neighbor rule to classify the sample type.
出处
《重庆科技学院学报(自然科学版)》
CAS
2010年第4期161-164,共4页
Journal of Chongqing University of Science and Technology:Natural Sciences Edition
关键词
模式分类
概率密度函数
非参数技术
pattern classification
probability density function
non-parametric technique