论文部分内容阅读
本文对乡镇级初级卫生保健的５２个总指标按效能设计出３３个派生指标，然后用变异系数法，聚类分析法，因子分析法三种方法从３３个派生指标中挑选具有代表性和特异性的１５个初选指标。用此１５个指标进行样本聚类，其聚类图同未筛选指标时的聚类图基本相似，反映出１５个指标保留了原３３个指标的大部分信息。根据样本聚类所得的类别进行逐步判别分析，筛选出最后的１０个指标。其判别函数经回代和新样品的验证，其判别函数国代率为９５．５２％，符合率为８０％。为了使那些用于评价区分性差而未进入判别函数的指标中某些重要指标不被忽视，又建立了含１５个指标和含２２个指标的判别函数，决策者可以根据实际情况，从上述三套判别函数中选择一套供基层自评用。
This article designs 33 derivative indicators based on 52 indicators of primary health care at the township level according to their performance, and then selects from 33 derived indicators using the three methods of variation coefficient method, cluster analysis method and factor analysis method to be representative and specific. Sexual 15 primary indicators. Using these 15 indicators for sample clustering, the clustering maps are basically similar to the clustering maps when no indicators were selected, reflecting that 15 indicators retained most of the original 33 indicators. Stepwise discriminant analysis was performed according to the categories obtained by clustering the samples, and the last 10 indicators were selected. After the discriminant function was verified by the back generation and new samples, the discriminant function country rate was 95.52%, and the coincidence rate was 80%. In order to make some of the important indicators used in the assessment of poor discrimination without entering the discriminant function not to be ignored, a discriminant function with 15 indicators and 22 indicators was established. The decision makers can use the above three criteria according to actual conditions. A set of discriminant functions is selected for self-evaluation by the grass-roots level.