| 摘要: |
| 针对传统的主成分分析与Topsis结合评价方法，会遇到提取的是变量的线性组合，损失了原始变量的初始值，且各主成分需根据变量的占有率高低重新命名，容易导致主成分含义不明显，指标解释能力不足的弊端，提出引入改进主成分分析法；采用改进后的主成分分析法从15个评价指标中提取出具有代表性的原始的9个指标，并根据指标的贡献率归一化客观确定其权重，构造加权规范矩阵与Topsis法结合，将该方法应用到学术期刊的评价中，得出结论与实际情况接近，说明该组合的评价方法具有一定的实用意义。 |
| 关键词: Topsis 改进主成分分析 权重 期刊评价 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Research on the Application of Topsis and Improved Principal Component Analysis to the Evaluation of Academic Periodicals |
|
ZHENG Hang
|
| Abstract: |
| The combination evaluation method of traditional principal component analysis and Topsis can lose the initial value of original variables because it meets the extraction of linear combination of the variables and because each principal component needs to be renamed according to its sharing level，and can cause the unclear meaning of the principal component and insufficient index explanation capacity，based on that，the improved principal component analysis is pointed out. The improved principal component analysis is used to extract representative original 9 indicators from 15 evaluation indicators，to objectively determine their weight according to contribution rate normalization，and to construct weighted normalized matrix and combine with Topsis. This method has been used in the evaluation on academic periodicals and the produced conclusion is close to practical situation，which reveals that this combined evaluation method is of certain practical significance. |
| Key words: Topsis improved principal component analysis weight journal evaluation |
