Study On The Evaluation Model Of DEA Cross-efficiency

Data envelopment analysis (DEA) was first proposed by Charnes, Cooper and Rhodesin1978, a "data-oriented" type of efficiency evaluation tool for the evaluation of a group ofsimilar entities. The Group of similar entities was called "decision-making units (DMUs)",which was evaluated as a subject by DEA model, meeting the structure of multiple inputand multiple output indicators, the structure made the application of DEA model had thecharacteristics of generality and flexibility. DEA model was a non-parametric method forefficiency considerations. Charnes, Cooper and Rhodes described DEA model as amathematical programming model, which provided empirical analysis for observing data.The model could calculate measurement tools directly related to modern economics, suchas the production function and efficiency of the production possibility set. In fact, DEAmodel was some kind of frontier theory, not the theory of central tendency in statistics.DEA model reflected decision-making units on the segmented linear hyperplane instead ofmultivariate statistical regression hyperplane. Due to such characteristics, DEA model gotrid of the assumptions of linear and non-linear relations in multivariate statisticalregression analysis, easily revealing the existence of the hidden relationships in the systeminside, which were not summarized through other evaluation methods.Academia considers there are two disadvantages in the existing classic DEA model,one is DEA model can only divide the evaluated units into effective units or non-effectiveunits, but can not put them in order, the other is the self-assessment multiplier system iseasy to exaggerate the strengths of some inputs and outputs and avoid the weaknesses ofother inputs and outputs in the evaluated units, leading to impractical efficiencymeasurement results. In response to these problems, Sexton et al proposed the crossefficiency evaluation method on the basis of the CCR model. The method used"self-evaluation " and " peer-evaluation" strategies to evaluate the relative efficiency of thedecision-making units, weakening the extremity of classic DEA model’s relative efficiencymeasurement conclusion due to the self-evaluation weights system, and evaluatingdecision-making units in order, so the method received great attention of the relevantexperts and scholars.There might be multiple optimal solutions in traditional DEA models, resulting in thecorresponding cross-efficiency value may not be unique, it had become the main challengeof the validity of the impact of cross-efficiency evaluation method. With the help of technique for order preference by similarity to an ideal solution (TOPSIS), this articlemade the gap between the ideal point and evaluated units as metrics through constructingthe ideal point, giving new secondary objective programming model. The model on the onehand solved problems of the instability of the cross efficiency evaluation value, on theother hand solved the problem of decision-makers’ selecting difficulty proposed by Wang.Aiming at Sexton et al stated the treatment, that cross-efficiency evaluation methodusing the same weights to aggregate and gain cross efficiency evaluation of evaluated units,Wu et al questioned and stated non-equal-weighted cross efficiency evaluation method toevaluate each cross efficiency of the evaluated units, which was based on the concepts ofcooperation alliance in game theory, nuclear solution and the Shapley value, Wang et albelieved that, on the one hand, this method had the characteristic of calculation complexity,and on the other hand, the problem of not gaining the only one result, so with the help ofYager’s ordered weighted averaging (OWA) theory, Wang et al stated cross efficiencyevaluation method based on OWA, this method selected corresponding integration weightsdepending on the decision-makers’ optimism level, then obtained decision-making units’cross efficiency evaluation value by weighting average of every cross efficiency of thedecision-making units. The existing DEA ranking method based on the cross-efficiencyevaluation strategy only defined cross efficiency evaluation value for the sort ofdecision-making unit. Aiming at the construction method of this concept, my article fromthe management definition of efficiency and multi-attribute decision theory these two partsrespectively, stated that the evaluation value not only did not meet the managementdefinition of efficiency, but also had no rational logical relationship with good and bad ofdecision-making units. To solve the above problems, this article put forward a new conceptof comprehensive relative efficiency of DEA, and based on the optimization theorydesigned measurement model which can calculate DEA comprehensive relative efficiency.The model did not require the introduction of the decision-makers’ preference judgments,which could adhere to the objective analysis of the orientation of the DEA as anonparametric method better. In addition, the comparison of the data analysis showed thatcontrast to existing methods, DEA comprehensive relative efficiency model showed strongcomparative advantage in the ability to kill rank reversal of the decision-making unit.This article gave a general description of DEA comprehensive scoring model,introduced non-weight constrained confidence region and weight constrained confidenceregion these two types of models separately, especially pointed out that in weight constrained confidence region model, in fact the weight of CK model had been determinedonce the interval was finalized, and its essence was the application of the cross-evaluationmethod, cross-evaluation matrix which was gained by cross evaluation satisfied thejudgment matrix characteristics of AHP/ANP methods. In principle of relative evaluationto judge the matrix structure, combining with cross-evaluation method proposed generalsteps of relative score ranking method. Receiving the ranking results of judgment matrixby AHP/ANP power method, and a specific numerical application was given to prove thefeasibility of this method.Every classic DEA model including the CCR model, their suitable objects were onlythe input and the output of a single production system, even if there were two or even moreproduction structures in the inspection production system, and classified the intermediateproducts as input or output arbitrarily in the process of model application. While most ofthe production processes in reality have at least two or multiple production sub-processes,arbitrarily classified them is not close to reality, and does not reflect the actual performancelevel of the evaluated units.This article selected the fusion of cross-efficiency and two-stage model as abreakthrough, in view of the problem of the cross-efficiency not unique would continue toappear in the two stages of the evaluation process, in reference to the previously proposedideal decision-making unit method, two-stage cross efficiency model was stated, analysisof examples showed that the new model was well recognized for meeting the two-phaseproduction structure case. Similar to cross-efficiency method of single-stage whichcontained the ideal decision-making unit, new two-stage cross-efficiency model not onlysolved the problem of cross-efficiency was not unique, but also insisted on the theoreticalbasis that DEA model was optimal to evaluate the decision-making units.The development of Chinese online publishing of scientific papers mode is still in thelower stage. This article argued that improving the interactive review session, giving fullplay of the online publishing academic enlightening could effectively improve onlinepublishing quality, so this article chose the establishment of an interactive academicplatform as a specific breakthrough to promote Chinese scientific papers online publishing,and provided implementation of the framework and implementation of the program.Concrete implementation of the framework and implementation of the program includedthe establishment of the interactive academic platform and platform module scoring systembased on the cross-efficiency method. Interactive academic platform would have some practical significance for scientific papers online publishing.Combined the two-phase cross-efficiency evaluation model to sort12nationaljoint-stock commercial banks. Based on the characteristic of efficiency values of sub-stagemodel, this article received effective and objective results from profitability and marketcapacity these two fields. The model not only gave out the objective ranking results ofcomprehensive relative efficiency of the national joint-stock banks, but also lists theefficiency of the two aspects covering the profitability and market capacity. Commercialbanks can improve and enhance the corresponding weaknesses depending on the rankresults and efficiency sort results, combining with its own developing advantages andobjectives.12national joint-stock commercial banks were evaluated by two-stage crossefficiency evaluation model and sample data, which provided excellent reference for theuse of new model and method of bank evaluation.