Partial Order Comprehensive Evaluation Method for Panel Data Processing
With the continuous improvement of the ways and means of data collection,the data of various indus-tries show the characteristics of panel data.Although the information of panel data is more abundant than that of cross-sectional data,it is difficult for traditional comprehensive evaluation methods to deal with such data.In this paper,the partial ordered set method is used to solve the problem of comprehensive evaluation of panel data.Compared with cross-sectional data,panel data adds a time dimension,which is the core of limiting the application of traditional decision evaluation.The idea of"dimensionality reduction"is adopted to transform the panel data in a traditional manageable way,that is,the panel data is"compressed"into cross-section data through time weight.According to the poset theory,the sequence of weights(weight space)is used to replace the exact weights to solve the problem that the time weights can not be assigned accurately.Through the matrix transformation of the indexes in different periods,the weight information is imported,and the panel data is converted into cross-section data.In response to the difficulty in assigning weights to indicators in cross-sectional data,the partial set method is applied again to matrix-process the indicator data and obtain the final result of scheme comparison,that is,the partial order Hasse diagram is used to express the comprehensive evaluation result of the panel data.As a mixed data of time series and cross-sectional data,panel data involves three dimensions:time,space,and index.By using the comprehensive evaluation method expressed by partial order,the key points are as follows:(1)In practical application,regardless of time or index weight,only the weight order of the index is needed.The information dimension of expert preference is integrated into the model,considering the diversifica-tion of information sources,giving full play to the characteristics of the subjective weighting method,and reali-zing the full integration of information and data.(2)The Hasse diagram expresses the comprehensive evaluation results of panel data,and the hierarchical clustering information between schemes is shown visually.Determinis-tic information and uncertain information can also be displayed through the Hasse diagram.The comparable relationship of schemes reflects the robustness of scheme comparison.As long as the weight order remains unchanged,no matter how the accurate weight changes,the scheme comparison will not change.(3)The full sort can be realized according to the Hasse diagram,and this kind of full sort contains probability information,so it is a full sort with more abundant information.The partial order method is used to solve the weight problem so that the traditional comprehensive evaluation model can comprehensively evaluate the panel data at a low cost after the partial order is upgraded.The partial order method is very different from the previous weight processing methods.The partial order method no longer restricts the parameters to a certain value but uses the weight space to express the weight.The weight space used includes the ownership weight under the given preference,and the evaluation result has better robustness and unity.Decision makers can construct weight space according to personal preferences and judgments,and then reflect the personal characteristics of decision makers.As long as the weight space of time and index is defined,the evaluation results can be expressed by partial order Hasse diagram,which can not only compare the advantages with disadvantages but also reflect the degree of robustness.Finally,through the panel data of the logistics industry in 14 cities in Liaoning province,we can see that the partial order comprehensive evaluation method can not only effectively deal with the panel data,but also has the characteristic functions of robustness and stratification.
partial order comprehensive evaluationpanel datatime weightHasse diagram
国家科技期刊平台
NSTL
万方数据
