Design of evaluation system for higher vocational institutions based on AHP hierarchical analysis

The development of higher vocational education is inseparable from a country's economic and social development and industrial structure development. Strengthening scientific evaluation of the development status of higher vocational education is an important prerequisite and foundation for scientifically promoting the development of higher vocational education. Educational evaluation is an effective scientific method for discovering the shortcomings of institutions' development and promoting their improvement. Based on this, it is important to construct a systematic, rich, measurable, forward-looking and robust as an evaluation index system. This paper uses a combination of subjective expert scoring and objective AHP hierarchical analysis to construct an evaluation system for higher vocational institutions, and comprehensively analyses the relative strength indicators and relative weakness indicators of each vocational institution, with a view to helping relevant institutions understand their overall development situation more comprehensively and objectively, and promoting the sustainable development of higher vocational education.


Introduction
As human socio-economic development enters a new stage, the refinement of social division of labour and the promotion of industrial collaboration models are increasingly appearing in different fields. Vocational education is a bridge between the training of talents and the demand for social skills, and assumes a more important task in promoting industrial upgrading and economic restructuring. "How to develop vocational education well" is inevitably dependent on educational evaluation as an important tool for conveying quality information and assessing development. Research on education evaluation has been increasing by scholars and experts in related fields. The focus of research is mainly on universities and the evaluation of their professional, research, teaching and management aspects. In the area of institutional evaluation, OCTAVIO proposes a visual data extraction method to analysis the developmental changes of institutions [1], and Samuel Gratzl proposes a method of comparing multi-attribute rankings to compare the rankings of universities in different years [2]. This paper uses AHP hierarchical analysis to discover and study the comprehensive strength of the collected higher vocational institutions, and further improve the evaluation system through significance and consistency tests, with a view to providing a scientific basis for the planning and construction of higher vocational institutions in the future.

The theory of evaluation system of higher education institutions
The evaluation system of higher education institutions established in this paper is analyzed by AHP (Analytic Hierarchy Process), which is a method combining subjective judgment and scientific calculation proposed by Professor Thomas L. Saaty of the University of Pittsburgh in the early 1970's [3]. AHP embodies the idea of decomposing and synthesizing, Representation of different relationships by stratifying and categorizing the problems analyzed and step-by-step [4]. The application of hierarchical analysis is divided into three main steps: the first step is to construct the index system; the second step is to calculate the index weights; and the third step is to calculate the data of each institution according to the weights, resulting in an evaluation score table [5].

Constructing a hierarchical evaluation model
According to the development characteristics of higher vocational institutions, reference to the research results of relevant aspects. According to the principles of scientificity, systematicity, comparability, and operability [6], the evaluation factors were selected. The scale of school operation B1, major and curriculum construction B2, faculty strength B3, talent cultivation B4, social service B5 and scientific research strength B6 are initially determined as the guideline layers of the evaluation system. To ensure the independence and objectivity of the evaluation factors, relevant experts were invited to score the 30 evaluation indicators in Table 1. The scoring is based on the importance scale 1-9 recommended by Saaty, as shown in table 2. Both indicators are equally important 3 x indicator 3 times more important than y indicator (marginally important) 5 7 9 2,4,6,8 decimal x indicator 5 times more important than y indicator (clearly important) x indicator 7 times more important than y for (strongly important) X indicator 9 times more important than y-indicator (extremely important) Importance indices between the above descriptions x indicator is less important than y indicator The experts judge the importance of all the indicators collected and give them a corresponding score. Of these, the very important ones are left in the indicator system with a greater weight and the less important ones are given a smaller weight. The final results are calculated according to the different scores.
Since the sample size is less than 30, the test is carried out using the t-distribution commonly used for small samples. First, because the overall variance is unknown, the entire data obeys a distribution with degrees of freedom of n-1 and the test statistic is = In equation (1), the overall mean was replaced using the sample variance s^2. Results showing significance levels much less than 0.05, it proves that parameters are statistically insignificant [7].all indicators in Table 2 are retained for the next step of the calculation.

Determination of weights and consistency tests for first-level indicators
In this paper, a hierarchical single ranking method is used to calculate the importance of all the first level indicators in the criterion level for the target. There are two ways to achieve this, and this paper uses the square root method for calculation. In the judgment matrix A constructed in the previous step, the Mth power of the product of each row is calculated to obtain an M-dimensional vector In Equation (2), denotes the vectors in the judgement matrix A and m is the total number To ensure that the data is reasonable, a consistency test needs to be performed on each matrix. Use equation (4) to find the maximum characteristic roots of the judgment matrix and the corresponding eigenvectors [8].
In Equation (4), is the largest characteristic root in the judgment matrix, and w is the weight value of the index after standardization [9]. Then the Consistency index is calculated as follows: Equation (6), the random consistency index (Random Index, RI) by looking up the table 3, we use Saaty simulated 50,000 times to obtain a table of random consistency values [10]. The judgment matrix of the first-order indicators in this paper is six, and the CR value calculated by bringing in the RI value is 0.093, which is less than 0.1. The results show that the degree of consistency of the judgment matrix A is considered to be within the tolerable range, and the weight vector calculation can be carried out using the eigenvectors of A. The final calculation results are shown in Table (2.3).

Weight determination and consistency test of secondary indicators
The weight determination method of the secondary indicators is the same as that of the primary indicators, using the hierarchical single ranking method to obtain the weights of each secondary indicator for the corresponding primary indicators respectively. The two-factor judgment matrix of the second-level indicators under each first-level indicator is constructed, and the CR value of the matrix is calculated to see if it meets the consistency requirements. In this paper, six CR values were obtained for the second-level indicators under the six first-level indicators, which were 0.0083, 0. 0042, 0. 0358, 0. 0858, 0. 0452 and 0.0653 respectively, all of which were less than 0.1 and met the consistency requirement. Finally, according to the weight relationship between the primary indicators and the total target, the weights of the secondary indicators were calculated relative to the total target, as shown in Table (

Analysis of Evaluation Results
The data in this paper comes from the reports provided by institutions in the Ministry of Education's statistical annual report, and the evaluation analysis of 52 double-high institutions. The top 10 institutions with high overall evaluation scores are shown in Table 3.1.
From Table 4 we can see that the development of vocational institutions on different evaluation indicators is uneven, and each institution has its strengths and weaknesses in different aspects. Taking the second-ranked institution as an example, we can find out by comparing the relevant information on the official website of this institution. This institution is a national key vocational institution and has been selected to participate in the national " High level development Program". Under the guidance of the program, the university has achieved remarkable results in terms of professional and curriculum development and talent training, indicating that it has achieved good results in teaching and learning, and the quality of employment of its students is high. However, when compared vertically with other institutions, it can be found that its score in social services is low, indicating that this school has a low contribution to social and technical training and occupational Science Popularization, and it has not effectively driven the development of regional vocational educationrelated economic industries.it also proves that the integration of industry and education and schoolenterprise cooperation needs to be strengthened. Other vocational institutions can also adopt this similar analysis method for interpretation and analysis.

Summary and outlook
In this paper, the method of AHP was used to evaluate the comprehensive development of vocational institutions. The indicators were selected using the expert scoring method to count the data involved in the relevant indicators. In order to ensure that the questionnaire has objectivity, a mean ttest was conducted on different questionnaires, and the significance level showed that the probability of falling into "I error" did not exceed the given level, and the indicators could be retained for calculation. In terms of weight determination, the square root method of the hierarchical single ranking approach was chosen to calculate the weights of the indicators in the criterion level relative to the target level separately and Weighting of indicators at sub-criterion level relative to indicators at guideline level. Then this paper calculates the weights of the second-level indicators relative to the total target according to the weights of the first-level indicators, and finally calculates the evaluation results of vocational institutions.
It should be noted that the indicators in this paper are selected with the help of experts and related scholars' scoring tables, and some more objective methods can be selected for analysis in the subsequent research. The weights of the secondary indicators in the sub-criteria level can also be calculated using the hierarchical total ranking method, which is not repeated here.