基于均值化主成分分析的雾霾环境分析与研究  

Haze environment analysis and research based on equalization of PCA

作　　者：李莉[1] 孙永霞[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《计算机应用研究》2015年第5期1373-1375,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(11072090)

摘　　要：由于雾霾天数对政府从宏观上把握空气状况、制定相应的政策更具有实际意义,所以选择年度雾霾天数而非具体每天的天气状况为研究对象,以期为政府控制雾霾提供可行性建议。从城市开发状况、地区环境、区域经济发展等角度出发,寻求影响雾霾天数的关键因素。对影响雾霾环境的指标数据进行主成分分析,同时改进了传统的主成分分析法在数据无量纲化方面存在的缺陷,采用了在主成分分析时对数据进行均值化,有效地解决了标准化方法带来的信息丢失问题。通过实例分析,对改进前后结果的差异进行了比较,其结果表明,改进后的主成分分析法选取的主成分数量减少,主成分累计贡献率也越高。Due to the haze days had more practical sense for the government to grasp the air conditions from the macro level to formulate corresponding policies,so selected annual haze days rather than specific weather of every day as the research object,in order to provide feasible suggestions for the government to control the haze weather. From the condition of urban development,regional environment and regional economic development perspective,to seek the key factors influencing the haze days.Processing the index data that affecting the haze environment by PCA（ principal component analysis）,at the same time,improved the traditional PCA in the aspect of data dimensionless defects,Using the mean of data preprocessing in PCA solvesd the problems of the information loss in standardization method effectively. Finally through an instance analysis,compared the difference of the results before and after improvement,the results showed that the principal component analysis to select the number of principal components to reduce,and the higher the principal component cumulative comtribution rate.

关 键 词：均值化 无量纲化 主成分分析 贡献率 雾霾环境 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP301.6[自动化与计算机技术—控制科学与工程]