摘要
KNN算法中各样本特征被视为等值权重,特征之间的关联因素没有考虑在分类算法中。为了解决此问题,提出一种基于皮尔森相似度和距离权重的改进KNN算法,根据训练样本和待分类样本计算皮尔森相似度和距离权重来判定特征和类别的相关度,且提出一种贡献率类别的判定方法。仿真结果表明,与KNN算法相比,提高了算法的分类精确度。
In the KNN algorithm,each sample feature is treated as an equivalent weight,and the correlation factors between the features are not considered in the classification algorithm.In order to solve this problem,an improved KNN algorithm based on Pearson similarity and distance weight is proposed.The Pearson similarity and distance weight are calculated according to the training samples and the samples to be classified to determine the correlation between features and categories,and a contribution is proposed.The method of determining the rate category.The simulation results show that compared with the KNN algorithm,the classification accuracy of the algorithm is improved.
作者
尹欢一
文志诚
马正见
YIN Huan-yi;WEN Zhi-cheng;MA Zheng-jian(School of Computer,Hunan University of Technology,Zhuzhou 412007,China)
出处
《电脑知识与技术》
2019年第9X期208-210,共3页
Computer Knowledge and Technology
关键词
K近邻
皮尔森相似度
距离权重
相关程度
K nearest neighbor
Pearson similarity
Distance weight
Correlation degree
