| 一种用于综合评价的主成分分析改进方法 |
| |
| 引用本文： | 高艳,于飞.一种用于综合评价的主成分分析改进方法[J].西安文理学院学报,2011,14(1):105-108. |
| |
| 作者姓名： | 高艳 于飞 |
| |
| 作者单位： | 哈尔滨工程大学,理学院,黑龙江,哈尔滨,150001 |
| |
| 摘 要： | 主成分分析是多元统计分析中的降维技术,在用于综合评价时,在不损失原有信息的基础上,主成分分析结果易受异常值的影响,分析结果稳定性差.针对该问题,文中提出一种改进的主成分分析方法,该方法先通过惯性系数加权的方式对原始指标进行分级优化,再利用均值化的思想对其进行处理.实验结果表明,该方法有效地弱化了异常值的影响,增强了分析结果的稳定性,同时具有良好的降维效果.
|
| 关 键 词： | 主成分分析 综合评价 异常值影响 均值化 惯性系数 |
A Modified Principal Component Analysis Algorithm For Comprehensive Evaluation |
| |
| GAO Yan,YU Fei.A Modified Principal Component Analysis Algorithm For Comprehensive Evaluation[J].Journal of Xi‘an University of Arts & Science:Natural Science Edition,2011,14(1):105-108. |
| |
| Authors: | GAO Yan YU Fei |
| |
| Institution: | (College of Science,Harbin Engineering University,Harbin 150001,China) |
| |
| Abstract: | Principal component analysis is one of the techniques of dimensionality reduction in the multivariate statistical analysis.When applied in comprehensive evaluation with no loss of original information,the principal component analysis algorithm is vulnerable to the abnormal value and the stability of its result can not be guaranteed.To address this problem,a modified algorithm was given in this study.In the new model,the original index values are graded and optimized with inertial coefficient and then processed with equalization.The experimental results showed that the new method was more stable and more effective in weakening the impact of the abnormal value.Meanwhile,it had a good effect on the dimension reduction. |
| |
| Keywords: | principal component analysis comprehensive evaluation impact of abnormal value equalization inertial coefficient |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
