Research of Education Evaluation Information Mining Technology Based on Gray Clustering Analysis and Fuzzy Evaluation Method

This paper has surveyed the education evaluation method and technology both at home and abroad, and studied on the question of education evaluation information's mining and synthesis processing. In view of the evaluation index system of high and secondary vocational education, on the basis of gray clustering analysis method this paper had established gray clustering model and applied fuzzy evaluation method to solve the question of education evaluation information synthesis processing. It provided a vastitude future of education evaluation information's mining and synthesis processing.


Introduction
Education evaluation is that according to particular education value and education object, using operable science artifice, make value judgment on education action, processing and result by systemic collecting, analyzing and arranging information data.The result of judgment will provide a basis for continuous self-improvement and education decision.The issue of education evaluation information processing and data mining is an important issue of education evaluation research The solution of this issue will not only promote education quality improvement but also provide valuable reference information for increasing university teaching innovation and implementing the quality education.At the abroad the theory and application of education evaluation had a quickly development from this century, but the method and theory of evaluation information's integrated processing and data mining was behindhand in development relative to education evaluation investigation.So far, it is three representative education evaluation patterns including object direction evaluation pattern, decision direction evaluation pattern, pluralistic evaluation pattern in the development of abroad education evaluation.The education evaluation pattern investigation noted very little in the method and theory of information integrated processing and data mining.Each education evaluation had a definite value orientation, and these value orientations or value judgment were dominated by the study and application of information integrated processing and data mining theory and method.
Education is a complex system, so the integrated evaluation of complex system was needed many index to weighing.A part of internal study on education evaluation and evaluation index system was based on analytical hierarchy process, and found more science evaluation index system.When more science evaluation index system was found, we must choose a more precisely computation method.Accordingly, the researcher who at abroad and home had pay more attention to how choose a more precisely computation method in the education evaluation processing.In 1982, the famous researcher named Julong Deng in our country had brought up the gray system theory.The research object of gray system theory was uncertainty system what is unknown part of information's small sample or poor information.
The gray clustering is the method that cluster several observation index or observation object into some definable classes by gray association matrix or gray number's whitely weight function.(Yannis Caloghirou. 1999) (Chen Z. 2001)(Qiu Jianrong, Zhang Xiaoping, Liu Hao, Wang Quanhai, Li Fan & Zeng Hancai. 2002)(Gu Zhaojun, He Xiaohui, Si Zhensheng & Fan Jingxin. 2007) A clustering can be considered as a set that observation was belonged same class.
For the set what is constituted by all clustering object, we need not cluster by clustering index but also evaluate all evaluation object in a whole.For using data mining technology to point out the problem in the education evaluation and found the way to solve problem, this paper applies grey clustering analytic method as a basis and integration use fuzzy evaluation method (Sadaaki Miyamoto. 1990) to solve integration information processing in the education evaluation index system, thereby promote education evaluation's networking and informationization.At the same time according to high and secondary vocational education talented people training work level evaluation project, the paper use grey clustering method and fuzzy evaluation method to research and practice education evaluation in the evaluation project.The paper aspire after practicalness, pertinency and realistically in the content, materialize advanced and multiformity in the method, reach after science, justice and directional in the result.The research and solution of aforementioned problem will come about active effect and action for our country education evaluation technology.

Education evaluation information integration processing and data mining
The process of education evaluation data mining and information integration had three basic steps as follows:

Construct education evaluation index system hierarchy model and confirm index weight
After survey, education evaluation index system hierarchy's general model was confirmed by expert discussion again and again as figure 1 show: We use AHP-GA method to confirm each index weight (also can be confirmed by experts researching).We assume that the weight vector what first index 1 2 , ,..., n B B B relative to total object A is ( , , , )

Construct evaluation value matrix
It is assumed that the number of evaluation person was p , namely 1, 2, t p = K .The number of evaluation object was q , namely 1, 2, s q = K .From above index system, we assume that the number of first index was n , the number of second index was m and the number of third index was m k .
The evaluation person evaluates some evaluation object by third index's evaluation rank standard.We assume that the grade what the evaluation person t evaluates on evaluation object s by third index's evaluation rank standard was ( ) 1 ( 1, 2, , ; , , ; 1, 2, , ; 1, 2, , ) K , so the evaluation object s 's evaluation value matrix  After colligating all grey class's evaluation object s toward third evaluation indexes, the grey evaluation weight vector By colligating grey evaluation weight ) ( s ije r in the all of second index i C 's third index, it can be found that grey evaluation weight matrix ) (s i R of evaluation object s 's second index i C toward each evaluation grey class is: 2.5 Compute grey integration evaluation vector For the evaluation object s , according to 1 2 ( , , , ) that layer D's relatively weight vector corresponding layer C's element i C and grey evaluation weight matrix of second index i C toward each evaluation grey class, it can be evaluated integration, and then it found that grey integration evaluation vector ( ) s i TT of evaluation object s 's second index i C as follows: 2.2.6 Compute grey integration evaluation vector For the evaluation object s , by colligating grey integration evaluation vector ( ) s i TT in the each second index i C , it can be found that grey evaluation weight matrix ( ) s i L of evaluation object s 's first index n B toward each evaluation grey class is: According to 1 2 ( , , , ) that layer C's relatively weight vector corresponding layer B's element n B and grey evaluation weight matrix of first index n B toward each evaluation grey class, it can be evaluated integration, and then it found that grey integration evaluation vector ( ) s i

LT
of evaluation object's first index n B as follows: 2.2.7 Compute clustering result.
For the evaluation object s , by colligating grey integration evaluation vector According to the weight vector and grey evaluation weight matrix ( ) s Z of evaluation object s , it can be clustered integration, and then it found that clustering result ( ) s X of evaluation object as follows:

Apply fuzzy evaluation to finding evaluation result
It uses each grey class's threshold as rank value to compute the integration evaluation value of each evaluation object.For example, the grey class one's threshold is 1 d , the grey class two's threshold is 2 d ,…, the grey class g's threshold is g d .So each grey class rank value vector is 1 2 ( , , , ) . The integration evaluation value It can be made an order by the value of

G
after computing each evaluation object's integration evaluation value ) (s G .

Demonstration test
We use the evaluation process of high and secondary vocational education talented people training work level evaluation index system as example to explain the process of education evaluation information integration processing and data mining.According to the model of figure 1, firstly we have to confirm evaluation index's weight.Secondly, it evaluated on school A by five experts (data in table 1), then it found evaluation value matrix as follows: 8 7.5 8.5 8 7 7 6.5 7 6 7 6 6.5 7 6 6 8 8.5 9 8 7.5 7.5 8 7 6 7 According to grade standard of high and secondary vocational education talented people training work level evaluation project, it can be confirmed that the evaluation grey number is 5 e = , the whitely weight function is 1 2 3 4 5 ( ), ( ), ( ), ( ), ( ) f x f x f x f x f x as follows: It can be found that school A's all second index toward each evaluation grey class Then, it can be found that grey evaluation weight matrix

.
The cn means that the number of layer C's element relative to element n B .And for the same reason, the weight vector what layer D's weight relative to layer C η which valuation object s belongs to evaluation grey class e as follows maintained grey e on evaluation object s by third evaluation indexes was:

Figure 1 .
Figure 1.Education evaluation system hierarchy diagram Yannis Caloghirou.(1999).Mutivariate analysis for assessment of corporate performance.the case of Greece.Operational Tools in the Management of financial risks.