论文部分内容阅读
首先将不同量刚的指标标准化。其次对53个评分指标进行主成分分析,得到10个主分量(综合指标),能够综合提取53个指标中的85%的信息,再以10个主分量得分作为新的指标施行聚类分析从而对601个乡(镇)分类,最后利用典型相关分析检查分类情况。本文提出的方法可以合理地利用原始指标提供的客观数量信息及相关性将调查数据科学分类,避免由于指标太多难以把握分类尺度而导致的主客观偏差,也可以用上述方法对其他评定结果进行检查校正。
First, standardize the amount of different indicators. Secondly, principal component analysis was performed on 53 rating indicators to obtain 10 principal components (comprehensive indicators), 85% of the 53 indicators could be comprehensively extracted, and 10 principal component scores were used as new indicators to conduct cluster analysis. Classification was made to 601 townships (towns), and the classification was finally checked using canonical correlation analysis. The method proposed in this paper can reasonably use the objective quantitative information provided by the original indicators and the relevance to classify the survey data scientifically, avoiding the subjective and objective biases caused by too many indicators that are difficult to grasp the classification scale, and can also use the above method to perform other assessment results. Check the calibration.