Based On The Jackknife And PA-LV Model For Comprehensive Evaluation

In the application of comprehensive assessment,the high-dimensional data can lead to much difficulties to deal with.If the variables are independent of each other,in order to reduce dimension we could filter several main variables;If there is significant correlation among variables,the variable synthesis can be used in the form of extracting latent variables,thus achieving the purpose of dimension reduction.In the process of decreasing dimensional,sample size and whether there is a partial correlation among the variable,which has a certain effect on the choice of decreasing dimensional method.This paper mainly focuses on the relatively small sample size,and not only correlation is existent,but also the partial correlation among variables is unnegligible.First,the observation variables were divided by the method of integrating Jackknife method with exploratory factor analysis,and the structure of potential factors is determined by the combination of subjective cognition.Then,contains latent variable path analysis(PA-LV)model is builded,in order to estimate the path coefficient,Jackknife method is combined with the related theory of the path analysis,and then to calculate the score and sort the latent factors..Next,examine the correlation among latent factors;If independent,complete comprehensive evaluation through empowerment;If relevant,it is similar to find the potential factor method to calculate the final score.Finally,the article uses the actual medical data to analysis,and compares with the results of Analytic Hierarchy Process(AHP)and PLS Path Analysis.