摘要
借助初步计算,探索选择适宜统计方法的依据。方法:采用几种统计方法分析出生缺陷的地区聚集性,通过分析结果间比较,找出简便、有效的衡量统计方法优劣的指标。结果:累计贡献率,多维标度法前4个特征向量为84.97%,主成分分析法前2个特征向量为73.22%。结论:凡涉及特征值的统计方法,可根据与特征值相应的累计贡献率大小进行统计方法选择。
To explore basis for selecting statistical methods with preliminary calculation. Methods: Statisticalmethods were applied for data, to analyse spatial clustering of birth defects, indicatrix to judge medical statistics which is very goodor bad can be found with comparision among results of preliminary calculation. Results: After analysing the data, total contributory rate of multivariate scale for four eigenvectors was 84. 93%, one of pricipal. component analysis for two eigenvectors was73.22%. Conclusion: In involving statistical methods of eigenvalue, statistical method that total contributory rate is large, thatis relative to eigenvalue, is selected.
出处
《现代预防医学》
CAS
1999年第3期302-303,共2页
Modern Preventive Medicine
基金
国家自然科学基金!39470644
关键词
特征值
选择统计方法
应用
卫生统计
Eigenvalue Selecting statistical method Application condition of statistical method