论文部分内容阅读
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。
With preliminary calculations, explore the basis for selecting the appropriate statistical method. Methods: Several statistical methods were used to analyze the regional clustering of birth defects. By comparing the results, a simple and effective index for measuring the merits of statistical methods was found. Results: The cumulative contribution rate, the first four feature vectors for the multidimensional scaling method was 84.97%, and the first two feature vectors for the principal component analysis method were 73.22%. Conclusion: For statistical methods involving eigenvalues, statistical methods may be selected based on the cumulative contribution rate corresponding to the eigenvalues.