论文部分内容阅读
将学术期刊多属性评价方法分为线性评价方法与非线性评价方法,然后提出了一种通用的对非线性评价方法选取进行检验的方法。其原理是,基于正向评价指标值增加,评价结果增加的标准,首先采用非线性评价方法进行评价,然后用标准化后的评价指标作为自变量,评价结果作为因变量,采用偏最小二乘法进行回归,如果回归系数出现负数,在排除评价指标选择失误的基础上,说明评价方法选取不当。偏最小二乘法既可以作为对非线性评价方法选取进行检验的一种有效方法,也可以作为筛选评价指标的一种有效方法。
The academic journals multi-attribute evaluation method is divided into linear evaluation method and non-linear evaluation method, and then a universal method for testing non-linear evaluation method is proposed. The principle is based on the positive evaluation index value increases, the evaluation results increase the standard, first of all non-linear evaluation method for evaluation, and then use the standardized evaluation index as an independent variable, the evaluation results as dependent variables, using partial least squares If the regression coefficient is negative, on the basis of excluding the selection error of the evaluation index, it indicates that the evaluation method is not properly selected. Partial least-squares method can not only be used as an effective method to test the selection of non-linear evaluation methods, but also can be used as an effective method of screening evaluation index.