Institute student classification culture quality evaluation system and method based on interval model

Abstract

The invention discloses a graduate student classification cultivation quality evaluation system and method based on an interval model, and relates to the technical field of education evaluation. The system comprises a model construction module, an acquisition processing module, a matching processing module, a comprehensive evaluation module and an evaluation analysis module. Classification cultivation quality, scientific research and practice achievements and the like are taken as core indexes, an index interval is determined by combining historical data and industry standards, weights are calculated through an expert scoring mean value, and subjectivity and one-sidedness of traditional evaluation are avoided; for missing values, similar sample mean values are adopted for filling, abnormal values are eliminated according to interval screening logic, the data quality is guaranteed, and the method is suitable for multi-college, multi-specialty and cross-period evaluation scenes; relative positions of indexes are dynamically calculated through interval matching, a hierarchical evaluation result is obtained in combination with a weighting algorithm, advantages and shortages are accurately positioned through transverse and longitudinal comparison, a reliable basis is provided for supervision of education management departments and self-improvement of colleges and universities, and classified cultivation quality of graduate students is promoted to be efficiently improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of educational evaluation, in particular to a postgraduate classified training quality evaluation system and method based on an interval model. BACKGROUND

[0002] At present, the scale of postgraduate education in China is continuously expanding, and the contradiction between the demand for high-level talent cultivation and quality guarantee is increasingly prominent. Scientific evaluation of postgraduate classified training quality has become the core demand of educational management departments and universities. The existing postgraduate classified training quality evaluation system has significant shortcomings. First, the evaluation index system is not systematic, and it mainly focuses on the explicit dimensions such as scientific research practice achievements and curriculum settings, and does not fully cover the implicit indicators such as students' innovation and practice ability and professional development potential, making it difficult to fully reflect the overall picture of educational quality. Second, the weight setting is highly subjective, and traditional methods mainly rely on experience-based valuation, without fully considering the differences between industry standards and universities, resulting in the influence of some key indicators (such as resource guarantee and student development) being weakened, and the objectivity of the evaluation results being limited.

[0003] In terms of evaluation methods, the current mainstream methods such as analytic hierarchy process (AHP) and fuzzy comprehensive evaluation method have obvious limitations. These methods mainly use fixed numerical values as evaluation benchmarks, ignoring the differences in school-running foundation and development stage of different universities and majors, and lack of consideration of the reasonable fluctuation range of indicators. For example, the same "number of published papers" indicator has significant differences in reasonable intervals between research-oriented universities and application-oriented universities, and fixed benchmarks may lead to distorted evaluation. At the same time, there are loopholes in the data processing process, and simple mean filling is often used to deal with missing values, and there is a lack of standardized processing procedures for abnormal values, further affecting the accuracy of the evaluation results.

[0004] Although foreign mature evaluation systems (such as the CWUR ranking in the United States and the REF evaluation in the United Kingdom) focus on multi-dimensional quantification, they are difficult to directly adapt to the national conditions of postgraduate education in China due to differences in education systems and training objectives. Domestic research has attempted to integrate various evaluation methods, but has not yet formed a unified index interval standard and weight calibration mechanism, resulting in a lack of horizontal and vertical comparability of evaluation results for different universities and different time periods, and failing to provide precise improvement direction for educational quality improvement.

[0005] With the deepening of the "Double First-Class" construction and the promotion of postgraduate classified development, traditional evaluation methods have been unable to meet the needs of fine quality control. SUMMARY

[0006] (I) Technical problems solved In view of the deficiencies of the prior art, the present application provides a postgraduate classified training quality evaluation system and method based on an interval model, which solves the problems mentioned in the background art.

[0007] (II) Technical solutions To achieve the above object, the present application is implemented by the following technical solutions: a postgraduate classified cultivation quality evaluation method based on interval model, which comprises the following steps: First step, interval model construction: Step 1.1, select classified cultivation quality, scientific research practice achievement, student development, and resource guarantee corresponding four types of data as first-level indicators; then set several second-level indicators under each type of first-level indicator; for each second-level indicator, based on the pre-collected historical data and related industry standards, comparison data between colleges and universities, determine its minimum value and maximum value; which are the lowest level value and the highest level value that the second-level indicator can reach within a reasonable range; Step 1.2, mark each second-level indicator as i, where i=1, 2, …n, n is the total number of second-level indicators; then construct the corresponding single indicator interval [a i ,b i ] for each second-level indicator; Step 1.3, for each second-level indicator, invite experts in the field of postgraduate education to score the weight, and mark the weight score of each expert as S i,k , where k=1, 2, …m, m is the number of experts; then calculate the average value of the expert weight score of each second-level indicator i, and mark it as the average value of the expert score SP i ; then calculate the sum of the average value of the expert score SP i of all second-level indicators, and mark it as ST: then divide the average value of the expert score SP i by the sum ST of the average value of the expert score of all second-level indicators, to obtain the weight coefficient β i corresponding to each second-level indicator i; Second step, data collection and processing: Different colleges and universities, different majors, and different time periods in the same college and university are taken as a plurality of objects to be evaluated, and a single object to be evaluated is marked as sample j, where j is the serial number distinguishing different samples; each second-level indicator is collected from the school teaching management system, scientific research management platform, student employment statistics system, and resource procurement and management account corresponding channels; then the data collected from different channels are cleaned to process missing values and abnormal values; Third step, interval matching processing: Each evaluation object is marked as j1, and then the pre-processed data of each second-level indicator i of the evaluation object is matched with the corresponding single indicator interval [a i ,b i ]; Fourth step, comprehensive evaluation calculation: The second-level index i corresponding to each evaluation object j1 is weighted; then based on the weighted processing result, the comprehensive value of the first-level index is determined; then based on the comprehensive value of the first-level index, and in combination with the preset weight coefficient corresponding to the first-level index, the comprehensive evaluation result of the graduate classified cultivation quality of the evaluation object j1 is determined; the evaluation object j1 includes a university, a major, and a time period of the same university; Step 5, evaluation result analysis: The calculated comprehensive evaluation result of the evaluation object j1 and the comprehensive value of each first-level index are compared horizontally and vertically.

[0008] As a further scheme of the present application, in step 1.1: The first-level index corresponding to the classified cultivation quality includes the second-level index of teacher teaching satisfaction and course rationality. The first-level index corresponding to the scientific research practice achievement includes the second-level index of paper publication quantity and scientific research project participation degree. The first-level index corresponding to the student development includes the second-level index of employment quality and further education rate. The first-level index corresponding to the resource guarantee includes the second-level index of the number of teachers and the allocation of teaching facilities.

[0009] As a further scheme of the present application, in step 1.2: i And b i The determination method is as follows: For the second-level index i, first, the original data of the second-level index i is screened to remove outliers, and the screening method is: The original data of the second-level index i is sorted in ascending order, and then the first and last 5% of the original data is removed; Then in the remaining original data, the minimum value is taken as a i , and the maximum value is taken as b i .

[0010] As a further scheme of the present application, in step 1.3, the score range is 0 to 10 points, and 10 points are the most important.

[0011] As a further scheme of the present application, for the missing value, when the data of the second-level index i in the sample j is missing, the data corresponding to the second-level index i is extracted from the samples of the same type as the sample j, and is recorded as d i，1 , d i，2 , …… d i，p ; Wherein, p represents the number of samples of the same type as sample j.

[0012] Then, through , the filling value dB i.

[0013] As a further aspect of the present invention, the outlier is handled by removing outliers according to the method of filtering the original data of the secondary index i.

[0014] As a further aspect of the present invention, the matching method is as follows: The preprocessed data of secondary indicator i is labeled as X. i,j1 ; If data X i,j1 In [a i ,b i Within, then through Calculate X i,j1 In the corresponding interval [a i ,b i The relative position r within ] i,j1 ; Where, r i,j1 The value range is between 0 and 1; When x i,j1 =a i At that time, r i,j1 =0 indicates that it is at the lower limit of the interval; When x i,j1 =b i At that time, r i,j1 =1 indicates that it is at the upper limit of the interval; If X i,j1 >bi, then through Calculate X i,j1 In the corresponding interval [a i ,b i The relative position r within ] i,j1 Where γ1 is the preset reward coefficient for exceeding the upper limit; If X i,j1 <ai, then through , where γ2 is the preset penalty coefficient for exceeding the lower limit.

[0015] As a further aspect of the present invention, the weighted processing formula is as follows: Calculate C i,j1 This is the weighted value of the secondary indicator i for the evaluation object j1.

[0016] As a further aspect of the present invention, the formula for calculating the comprehensive value of the primary indicator is as follows: ; In the formula, C(A) j1 Let A be the comprehensive value of the primary indicator, and let A be a variable, where A ∈ [Category of training quality, scientific research practice results, student development, resource guarantee], and t(A) be the number of secondary indicators included in the primary indicator A, which is also a variable. 1,j1 C2,j1 , … C t（A），j1 is the weighted value of the secondary index i contained in the first index A.

[0017] As a further scheme of the present application, the calculation formula of the comprehensive evaluation result of the classified cultivation quality of postgraduates of the evaluation object j1 is , Q j1 is the comprehensive evaluation result of the classified cultivation quality of postgraduates of the evaluation object j1, and α(A) represents the preset weight coefficient corresponding to the different first index A. represents the sum of the product of the comprehensive value of the classified cultivation quality, the scientific research and practice achievement, the student development and the resource guarantee, respectively, and the preset weight coefficient corresponding to the classified cultivation quality, the scientific research and practice achievement, the student development and the resource guarantee, respectively.

[0018] As a further scheme of the present application, the horizontal comparative analysis is to compare the Q j1 and C(A) j1 of different evaluation objects, find out the advantages and gaps, and the specific steps are as follows: Two universities are selected as the evaluation objects, and the Q j1 of the university 1 and the university 2 are respectively recorded as Q1 and Q2, and then Q1 and Q2 are compared: If Q1>Q2, then the comprehensive values of the university 1 and the university 2 corresponding to the first index A are extracted, and are respectively marked as C(A)1 and C(A)2; When C(A)1>C(A)2, it indicates that the university 1 is stronger than the university 2 in the corresponding first index of the classified cultivation quality or the scientific research and practice achievement or the student development or the resource guarantee; and if C(A)1 is higher than the corresponding preset first comprehensive threshold C(A) max , it indicates that the university 1 is good in the corresponding first index A as a whole, and if C(A)2 is lower than the corresponding preset second comprehensive threshold C(A) min , it indicates that there is a problem in the corresponding first index A of the university 2.

[0019] As a further scheme of the present application, the longitudinal comparative analysis is to compare the evaluation results of the same evaluation object in different time periods, analyze the development trend of the classified cultivation quality of postgraduates, and the specific steps are as follows: Firstly, the Q j1 corresponding to e time periods of a university is selected, and then the Q j1 corresponding to e time periods of the university are recorded as Q1, Q2, … Q e according to the time trend; Then, the difference value QC r of adjacent two time periods is calculated through , wherein r=1, 2, … e-1, and e-1 represents the number of the difference values of adjacent two time periods. At the same time Calculate Q1, Q2, ... Q e The average value of QP; Then, the difference QC between each two adjacent time periods is calculated. r The difference threshold QC between the two values ​​is... y Comparison: When QC r >QC y When the time interval is r+1, it indicates that the time interval is improved relative to the time interval r, and the period from the time interval r to the time interval r+1 is recorded as the improvement period; When QC r <-QC y When the time interval is r+1, it indicates that the time interval is decreasing relative to the time interval r, and the period from the time interval r to the time interval r+1 is recorded as the decreasing period; When -QC y ≤QC r ≤QC y When the time interval is r+1, it indicates that the time interval is relatively stable compared to the time interval r. Then, extract the number of rising and falling periods respectively, and denot them as TS and TJ respectively, and calculate the percentage of TS and TJ in e-1 respectively; At the same time, Q1 and Q e Comparison with QP: when >60%, Q e >Q1 and Q e >If QP is true, then it is determined that there is an overall improvement in e time periods; when >60%, Q e If both Q1 and QP are true, then it is determined that there is an overall decline in e time periods; If it is determined that there is no overall increase or decrease, it means that there is overall stability in the e time periods.

[0020] A postgraduate classification and training quality evaluation system based on an interval model, the system being used to implement a postgraduate classification and training quality evaluation method based on an interval model, the system comprising: Model construction module: Selecting categorized training quality, scientific research achievements, student development, and resource support as primary indicators; establishing several secondary indicators under each primary indicator; for each secondary indicator, determining its minimum and maximum values ​​based on pre-collected historical data and relevant industry standards and inter-university comparative data; labeling each secondary indicator as i, where i = 1, 2, ..., n, where n is the total number of secondary indicators; subsequently constructing a corresponding single indicator interval [a] for each secondary indicator. i ,b i]; for each secondary indicator, invite experts in the field of graduate education to score the weight; then calculate the average value of the expert weight score of each secondary indicator i, and record it as the average value of the expert score; then calculate the sum of the average value of the expert score corresponding to all secondary indicators; by dividing the average value of the expert score by the sum of the average value of the expert score corresponding to all secondary indicators, the weight coefficient β corresponding to each secondary indicator i is obtained i ; The collection processing module: taking different colleges and universities, different majors and different time periods in the same college and university as a plurality of to-be-evaluated objects, a single to-be-evaluated object is denoted as sample j, wherein j is a serial number for distinguishing different samples; missing values and abnormal values are processed by collecting each secondary indicator from different channels and cleaning them; The matching processing module: each evaluation object is denoted as j1, j1 includes colleges and universities, majors and time periods in the same college and university, and then the preprocessed data of each secondary indicator i of j1 is matched with the corresponding single indicator interval [a i ,b i ]; The comprehensive evaluation module: each secondary indicator i corresponding to each evaluation object j1 is weighted; then the comprehensive value of the primary indicator is determined based on the weighted processing result; then the comprehensive evaluation result of the graduate classified training quality of the evaluation object j1 is determined based on the comprehensive value of the primary indicator and in combination with the preset weight coefficient corresponding to the primary indicator; The evaluation analysis module: the comprehensive evaluation result of the evaluation object j1 and the comprehensive value of each primary indicator are compared and analyzed horizontally and vertically.

[0021] (Three) beneficial effects The present application provides a graduate classified training quality evaluation system and method based on an interval model. Compared with the prior art, the following beneficial effects are achieved: In the present application, the mean value of expert scoring and the weight coefficient are calculated to avoid single subjective judgment bias; at the same time, the secondary indicator interval [a i ,b i ] is determined based on historical data and industry standards, the maximum value is taken after removing the first and last 5% abnormal values, the interval range is ensured to fit the actual level, a scientific foundation is laid for subsequent evaluation, and the problems of one-sided traditional evaluation indicators and subjective weight are solved.

[0022] In the present application, the data collection covers different colleges and universities, different majors and different time periods in the same college and university, and the sample dimension is rich; for missing values, the same sample data mean is used for filling, and the abnormal values are removed according to the index interval screening logic, which ensures the data integrity and accuracy. Compared with the problem that missing values and abnormal values are easily ignored in traditional data processing, this method reduces the interference of data errors on the evaluation result through the standardized cleaning process, makes the evaluation conclusion more reliable, and is suitable for multi-scenario graduate classified training quality evaluation.

[0023] The application realizes the dynamic matching of index data and interval by relative position calculation, sets reward coefficient for data exceeding the upper limit of the interval and penalty coefficient for data below the lower limit, avoids the rigidity of fixed threshold; the first-level index and the comprehensive evaluation result are calculated by weighting layer by layer, which can reflect the contribution of each dimension and form the overall evaluation. The method can accurately distinguish the advantages and gaps of different evaluation objects and provide a clear direction for targeted improvement of colleges and universities.

[0024] The application can intuitively compare the comprehensive and single index levels of different colleges and universities through horizontal comparison, and judge the problem short board combined with the threshold; the quality development and change of the same college are tracked through period difference and trend analysis in longitudinal comparison, and the two kinds of analysis form a three-dimensional evaluation perspective. At the same time, the supporting system module realizes the landing of the method, which can meet the needs of education management department supervision, college self-diagnosis, professional construction evaluation and other scenes, and help the precise improvement of classified cultivation quality of postgraduates. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 It is a system block diagram of a postgraduate classified cultivation quality evaluation system based on an interval model.

[0026] Figure 2 It is a flowchart of a postgraduate classified cultivation quality evaluation method based on an interval model. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0028] Please refer to Figure 1 and Figure 2 The embodiments of the application provide the following technical solutions: As an embodiment of the application: The application is a postgraduate classified cultivation quality evaluation method based on an interval model, comprising the following steps: First step, interval model construction: Step 1.1, interval construction preparation: First, select classified cultivation quality, scientific research and practice achievements, student development and resource guarantee corresponding to four types of data as first-level indexes; Then, set a plurality of second-level indexes under each type of first-level index; The primary indicators corresponding to the quality of categorized training include secondary indicators: teacher satisfaction and curriculum rationality. The primary indicators corresponding to scientific research practice achievements include secondary indicators: number of published papers and participation in scientific research projects. The primary indicators for student development include the secondary indicators: employment quality and further education rate. The primary indicators for resource guarantees include secondary indicators: number of teachers and availability of teaching facilities. For each secondary indicator, its minimum and maximum values ​​are determined based on pre-collected historical data, relevant industry standards, and inter-university comparative data. In this embodiment, a secondary indicator is selected and labeled X. Then, through statistical analysis of data from several universities and several time periods, the minimum value X of the secondary indicator is determined. min and maximum value X max , where X min X refers to the lowest possible level of the secondary indicator that can be achieved within a reasonable range. max It refers to the highest level value that the secondary indicator can achieve within a reasonable range; Step 1.2, Determining the single indicator range: For each secondary indicator, denote it as variable i, where i = 1, 2, ..., n, and n is the total number of secondary indicators; Subsequently, a corresponding single-indicator interval is constructed for each secondary indicator [a] i ,b i ], a i and b i The determination method is as follows: For secondary indicator i, the raw data for secondary indicator i is first filtered to remove outliers. The filtering method is as follows: The original data of secondary indicator i are sorted in ascending order, and then the first and last 5% of the original data are removed respectively. In this embodiment, the first and last 5% of the original data are rounded down. That is, assuming there are 30 original data for secondary indicator i, and after sorting, the first and last 5% of the original data each account for 1.5. The rounding process is to remove the extreme values ​​of the first 2 and last 2 original data after sorting, and then retain 26 original data. Then, from the remaining original data, take the minimum value as 'a'. i Take the maximum value as b i ; In this embodiment, it is assumed that: for the secondary indicator corresponding to teacher teaching satisfaction, data from 50 universities were collected. After sorting, the top 3 and bottom 3 extreme values ​​were removed, leaving 6. Among the remaining 44 data points, the minimum value among them is the lower limit 'a' of the single indicator interval. iwherein the maximum value is the upper limit value b of the single index interval i ; Step 1.3, index weight determination: For each secondary index, invite experts in the field of graduate education to score the weight, wherein the score range is 0 to 10 points, and 10 points is the most important, and the number of experts is denoted as m, and in this embodiment, m≥10; For secondary index i, the weight score of the kth expert is denoted as S i,k , wherein k = 1, 2, … m; The average value SP of the expert weight score of each secondary index i is calculated by i , which is denoted as the expert score average value; the expert score average value SP i indicates the importance of secondary index i in the judgment of experts; Then the sum ST of the expert score average values SP of all secondary indexes is calculated by i ; The weight coefficient β of each secondary index i is determined by i ; In this embodiment, the weight coefficient β i indicates the weight of secondary index i in the entire evaluation system, and the sum of all β i is 1, which is used for weighting different indexes in subsequent comprehensive evaluation; Step 2, data collection and processing: In the process of evaluating the quality of classified cultivation of postgraduates, data of different universities, different majors and different time periods of the same university are taken as multiple objects to be evaluated, and a single object to be evaluated is denoted as sample j, wherein j is the serial number distinguishing different samples; First, for each selected secondary index, data is collected from the school's teaching management system, scientific research management platform, student employment statistics system, and resource procurement and management account corresponding channels; In this embodiment, for example, for the teacher teaching satisfaction, data is obtained through the student online teaching evaluation system at the end of each semester; for the scientific research project participation degree, data is extracted from the project declaration and participation records of the scientific research department.

[0029] Then the data collected from different channels is cleaned to process missing values and outliers; For missing values, when the data of secondary index i in sample j is missing, the corresponding data of secondary index i is extracted from samples of the same type as sample j, and is denoted as d i，1 , d i，2 , … d i，pWhere p represents the number of samples of the same type as sample j; Subsequently passed Calculate the imputation value (dB) for the missing data of secondary indicator i in sample j. i ; For outliers, remove them using the same method as filtering the original data of secondary indicator i. Step 3: Interval matching processing: For each evaluation object, preprocess the data of its various secondary indicators i and their corresponding single indicator intervals [a] are compared. i ,b i Matching; The preprocessed data of secondary indicator i is labeled as X. i,j1 Where j1 represents the evaluation object; If data X i,j1 In [a i ,b i Within, then through Calculate X i,j1 In the corresponding interval [a i ,b i The relative position r within ] i,j1 ; Where, r i,j1 The value range is between 0 and 1; When x i,j1 =a i At that time, r i,j1 =0 indicates that it is at the lower limit of the interval; When x i,j1 =b i At that time, r i,j1 =1 indicates that it is at the upper limit of the interval; If X i,j1 >bi, then through Calculate X i,j1 In the corresponding interval [a i ,b i The relative position r within ] i,j1 Wherein, γ1 is the preset reward coefficient for exceeding the upper limit, and the value of γ1 is 0.5; If X i,j1 <ai, then through Where γ2 is the preset penalty coefficient for exceeding the lower limit, and the value of γ2 is 0.5; Step 4: Comprehensive Evaluation Calculation For each evaluation object j1, through The weighted value of the secondary indicator i of the evaluation object j1 is calculated as C. i,j1 ; Meanwhile, for each type of first-level index, it is marked as a variable A, wherein A is in [classified training quality, scientific research practice achievement, student development, resource guarantee], and then the number of second-level indexes contained by the first-level index A is extracted, wherein the number of second-level indexes contained by different first-level indexes is different; then the number of second-level indexes contained by different first-level indexes A is marked as a variable t(A), and the weighted value of the second-level index i contained by the first-level index A is marked as C 1,j1 , 2,j1 ,……C t（A），j1 ; Then, the comprehensive value C(A) of the first-level index A is calculated through j1 ; Wherein, C(A) j1 The comprehensive value of classified training quality: C(classified training quality) j1 The comprehensive value of scientific research practice achievement: C(scientific research practice achievement) j1 The comprehensive value of student development: C(student development) j1 The comprehensive value of resource guarantee: C(resource guarantee) j1 ; Then, the comprehensive evaluation result Q of the classified training quality of the graduate student evaluation object j1 is calculated through j1 ; Wherein, alpha(A) refers to a preset weight coefficient corresponding to different first-level indexes A, and alpha(A) includes the weight coefficient of classified training quality: alpha(classified training quality), the weight coefficient of scientific research practice achievement: alpha(scientific research practice achievement), the weight coefficient of student development: alpha(student development), and the weight coefficient of resource guarantee: alpha(resource guarantee), Refers to the sum of the products of the comprehensive values of classified training quality, scientific research practice achievement, student development and resource guarantee and the preset weight coefficients corresponding to classified training quality, scientific research practice achievement, student development and resource guarantee, respectively. The embodiment determines the weight by relying on expert scoring, and then obtains the comprehensive evaluation result through interval matching and weighted calculation. The whole process is logically rigorous, the indicators are selected comprehensively and the weight distribution is professional, which can objectively reflect the classified training quality of the graduate student, provide a scientific and standardized basic method for subsequent evaluation, avoid the deviation of single index or subjective judgment, and improve the evaluation credibility.

[0030] As the second embodiment of the present application: ​In the specific implementation, compared with the first embodiment, the technical solution of the embodiment only differs from the first embodiment in that the embodiment further comprises an evaluation result analysis step: the step compares the comprehensive evaluation result Q j1 of the evaluation object j1 and the comprehensive value C(A) of each first-level index j1 to perform horizontal and vertical comparative analysis. The evaluation object j1 comprises a university, a major, and time period data of the same university. The horizontal comparative analysis is to compare the Q j1 of different evaluation objects and the C(A) j1 to find out the advantages and gaps, as follows: Taking two universities as the evaluation objects as an example, the Q j1 of the university 1 and the university 2 are denoted as Q1 and Q2 respectively, and then Q1 and Q2 are compared. If Q1>Q2, then the comprehensive values of the university 1 and the university 2 corresponding to the first-level index A are extracted and denoted as C(A)1 and C(A)2 respectively. When C(A)1>C(A)2, it indicates that the university 1 is stronger than the university 2 in the corresponding first-level index A of classified cultivation quality or scientific research and practice achievement or student development or resource guarantee. If C(A)1 is higher than a first comprehensive threshold value C(A) max corresponding to the first-level index A, it indicates that the university 1 is good in the corresponding first-level index A, and if C(A)2 is lower than a second comprehensive threshold value C(A) min corresponding to the first-level index A, it indicates that the university 2 has problems in the corresponding first-level index A. In the embodiment: When C(classified cultivation quality)1>C(classified cultivation quality)2, it indicates that the university 1 is stronger than the university 2 in the classified cultivation quality. If C(classified cultivation quality)1 is higher than a first comprehensive threshold value C(classified cultivation quality) max corresponding to the classified cultivation quality, it indicates that the university 1 is good in the teaching-related links corresponding to the teacher teaching and course setting, and if C(classified cultivation quality)2 is lower than a second comprehensive threshold value C(classified cultivation quality) min corresponding to the classified cultivation quality, it indicates that the university 2 has many problems in the teaching links, such as the teacher teaching is not popular and the course is old and unreasonable. When C(scientific research and practice achievement)1>C(scientific research and practice achievement)2, it indicates that the university 1 is stronger than the university 2 in the scientific research and practice achievement. If C(scientific research and practice achievement)1 is higher than a first comprehensive threshold value C(scientific research and practice achievement) max, it indicates that University 1 is outstanding in aspects such as scientific research output and achievement transformation, while C (scientific research practice achievements) 2 is lower than the pre-set second comprehensive threshold C (scientific research practice achievements). min , it indicates that the scientific research training link of University 2 is weak, which may be caused by problems such as few opportunities for students to participate in scientific research and difficulties in achieving results. When C (student development) 1 > C (student development) 2, it means that University 1 is stronger than University 2 in student development. If C (student development) 1 is higher than the pre-set first comprehensive threshold C (student development) max , it indicates that University 1 performs well in the development paths of students such as employment and further study, while C (student development) 2 is lower than the pre-set second comprehensive threshold C (student development). min , it reflects that the development of students after graduation encounters obstacles, indicating that University 2 needs to strengthen in aspects such as employment guidance and further study support. When C (resource guarantee) 1 > C (resource guarantee) 2, it means that University 1 is stronger than University 2 in resource guarantee. If C (resource guarantee) 1 is higher than the pre-set first comprehensive threshold C (resource guarantee) max , it indicates that the resource allocation such as teachers and facilities corresponding to University 1 can effectively support education, while C (resource guarantee) 2 is lower than the pre-set second comprehensive threshold C (resource guarantee). min , it indicates that the resources of University 2 are insufficient, such as shortage of teachers and old equipment, which will limit the improvement of education quality. Vertical comparative analysis is to compare the evaluation results of the same evaluation object at different time periods and analyze the development trend of the classified cultivation quality of its postgraduate students, as follows: First, select the Q corresponding to a university in e time periods j1 as an example, and then record the Q corresponding to this university in e time periods according to the time trend j1 as Q1, Q2,... Q e ; Next, through , calculate the difference QC between two adjacent time periods r , where r = 1, 2,... e - 1, and e - 1 refers to the number of differences corresponding to two adjacent time periods; At the same time, through , calculate the average value QP of Q1, Q2,... Q e ; After that, compare the differences QC between each two adjacent time periods r with the pre-set difference threshold QC y : When QC r > QC yWhen Qr+1-Qr>0, it means that the r+1 period is improved relative to the r period, and the r period to the r+1 period is recorded as an improved period; When QC r <QC y When Qr+1-Qr<0, it means that the r+1 period is decreased relative to the r period, and the r period to the r+1 period is recorded as a decreased period; When -QC y ≤QC r ≤QC y When Qr+1-Qr=0, it means that the r+1 period is stable relative to the r period; Then, the number of improved periods and decreased periods is extracted respectively, and recorded as TS and TJ respectively, and the percentage of TS and TJ in e-1 is calculated; Meanwhile, Q1, Q e and QP are compared: When Q1 >60%, Q e >Q1 and Q e >QP are all true, it is judged that there is an overall improvement in the e periods; When Q1 >60%, Q e <Q1 and Q1 When it is judged that there is no overall improvement and overall decrease, it means that there is an overall stability in the e periods; The embodiment increases the evaluation result analysis on the basis of embodiment one, and compares the comprehensive value and the first-level index value of different objects horizontally, so that the advantage gap and problem index can be accurately positioned; the period data of the same object is analyzed longitudinally, and the development trend is judged in combination with the difference value and the average value. The threshold value is used to assist in judging the problem root, such as the deficiency of the teaching or scientific research link of the university, so that the evaluation can be extended from "resulting" to "promoting", and the practical guiding significance of the evaluation is enhanced.

[0031] As an embodiment of the present application: In the specific implementation of the present application, compared with embodiment one and embodiment two, the technical scheme of the present embodiment is to combine the schemes of embodiment one and embodiment two; The embodiment combines the scientific evaluation method of embodiment one and the result analysis scheme of embodiment two, which not only guarantees the objectivity of the evaluation itself by means of the interval model, professional weight and rigorous data processing, but also excavates the quality difference and development trend through horizontal and longitudinal comparison analysis, avoids the limitation of single embodiment, makes the evaluation system more complete, can not only output reliable evaluation results, but also provide comprehensive basis for optimizing the quality of postgraduate classified cultivation, and improves the comprehensive application value of the scheme.

[0032] The application scenarios of the present application can be as follows: According to the analysis results: For the management department of the university, resource adjustment and policy optimization can be targeted; If it is found that the comprehensive value of the scientific research and practice of a certain university is low, and through analysis, the number of published papers is lagging behind, policies can be introduced to encourage scientific research paper writing and publication, such as providing more scientific research guidance and increasing scientific research funding support; For the education department, the evaluation results of different universities can be used for classified management and guidance. More resources and policy pilot opportunities can be given to universities with high education quality, and improvement suggestions and assistance measures can be proposed to universities with quality to be improved; For students and parents, the evaluation results can be used as a reference for choosing universities and majors, understanding the actual situation of graduate classification training quality in different universities, and making more reasonable decisions.

[0033] It should be noted that all user data collected in this application is collected with the consent and authorization of the user, and the use of user data is legal and compliant, and the use and processing of user data comply with relevant laws, regulations and standards in the relevant region.

[0034] Meanwhile, the contents not described in detail in the specification are all the existing technology known to those skilled in the art.

[0035] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0036] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

[0037] Having described various embodiments of the application, it is to be understood that the above description is meant not to limit and not to encompass all of the possible embodiments. Many modifications and variations of this application can be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. It is intended that the scope of the application be defined by the scope of the patent and by the claims as allowed by the patent office, which can include adaptations and modifications. It is further intended that each of the individual elements or variations of this application be deemed to be disclosed herein.

Claims

1. A method for evaluating the quality of postgraduate training based on an interval model, characterized in that, The method includes the following steps: Step 1: Interval Model Construction: Select the following primary indicators: quality of categorized training, research practice achievements, student development, and resource support. Establish several secondary indicators under each primary indicator. For each secondary indicator, determine its minimum and maximum values ​​based on pre-collected historical data, relevant industry standards, and inter-university comparative data. Label each secondary indicator as i, where i = 1, 2, ..., n, where n is the total number of secondary indicators. Then, construct a corresponding single-indicator interval [a...] for each secondary indicator. i ,b i For each secondary indicator, experts in the field of postgraduate education are invited to assign weighted scores. Then, the average expert weighted scores for each secondary indicator i are calculated and recorded as the average expert score. Next, the sum of the average expert scores for all secondary indicators is calculated. The weight coefficient β for each secondary indicator i is obtained by dividing the average expert score by the sum of the average expert scores for all secondary indicators. i ; The second step is data collection and processing: different universities, different majors, and different time periods of the same university are used as multiple objects to be evaluated. Each object to be evaluated is recorded as sample j, where j is the sequence number that distinguishes different samples; various secondary indicators are collected from different channels and cleaned to handle missing values ​​and outliers. Step 3, Interval Matching Processing: Each evaluation object is denoted as j1, where j1 includes the university, major, and time period within the same university. Then, the preprocessed data of each secondary indicator i is matched with its corresponding single indicator interval [a...]. i ,b i Matching; Step 4: Comprehensive evaluation calculation: The secondary indicator i corresponding to each evaluation object j1 is weighted; then, based on the weighting result, the comprehensive value of the primary indicator is determined; then, based on the comprehensive value of the primary indicator and combined with the preset weight coefficients corresponding to the primary indicator, the comprehensive evaluation result of the postgraduate classification training quality of evaluation object j1 is determined. Step 5: Evaluation Result Analysis: Conduct horizontal and vertical comparative analysis on the comprehensive evaluation results of the calculated evaluation object j1 and the comprehensive values ​​of each primary indicator.

2. The method for evaluating the quality of postgraduate training based on an interval model according to claim 1, characterized in that: [the method is described in the original text]. The primary indicators corresponding to the quality of categorized training include secondary indicators: teacher satisfaction and curriculum rationality. The primary indicators corresponding to scientific research practice achievements include secondary indicators: number of published papers, research projects, and participation in practical internships; The primary indicators for student development include secondary indicators: employment quality and further education rate. The primary indicators for resource guarantees include secondary indicators: number of teachers, teaching facilities, and resources for school-enterprise cooperation. a i and b i The determination method is as follows: For secondary indicator i, the raw data for secondary indicator i is first filtered to remove outliers. The filtering method is as follows: The original data of secondary indicator i are sorted in ascending order, and then the first and last 5% of the original data are removed respectively. Then, from the remaining original data, take the minimum value as 'a'. i Take the maximum value as b i ; The scoring range is from 0 to 10, with 10 being the most important.

3. The method for evaluating the quality of postgraduate training based on an interval model according to claim 1, characterized in that: The cleaning method is as follows: For missing values: When the data of secondary indicator i is missing in sample j, the data corresponding to secondary indicator i is extracted from the samples of the same type as sample j, and its average value is calculated. Then, it is used as the imputation value for the missing data of secondary indicator i in sample j. For outliers: Remove outliers using the same method as filtering the original data for secondary indicator i.

4. The method for evaluating the quality of postgraduate training based on an interval model according to claim 1, characterized in that: The matching method is as follows: The preprocessed data of secondary indicator i is labeled as X. i,j1 ; If data X i,j1 In [a i ,b i Within, then through Calculate X i,j1 In the corresponding interval [a i ,b i The relative position r within ] i,j1 ; Where, r i,j1 The value range is between 0 and 1; When x i,j1 =a i At that time, r i,j1 =0 indicates that it is at the lower limit of the interval; When x i,j1 =b i At that time, r i,j1 =1 indicates that it is at the upper limit of the interval; If X i,j1 >b i Then through Calculate X i,j1 In the corresponding interval [a i ,b i The relative position r within ] i,j1 Where γ1 is the preset reward coefficient for exceeding the upper limit; If X i,j1 <ai, then through , where γ2 is the preset penalty coefficient for exceeding the lower limit.

5. The method for evaluating the quality of postgraduate training based on an interval model according to claim 4, characterized in that: The weighted processing formula is Calculate C i,j1 This is the weighted value of the secondary indicator i for the evaluation object j1.

6. The method for evaluating the quality of postgraduate training based on an interval model according to claim 5, characterized in that: The composite value of the primary indicator is the sum of the weighted values ​​of the primary indicator including the secondary indicator i, and is denoted as C(A). j1 ; Where A is a primary indicator, which is a variable, and A∈[Classification of training quality, scientific research practice results, student development, and resource guarantee].

7. The method for evaluating the quality of postgraduate training based on an interval model according to claim 6, characterized in that: The formula for calculating the comprehensive evaluation result of the quality of postgraduate training by category is as follows: Q j1 For the comprehensive evaluation result of the postgraduate training quality of evaluation object j1, α(A) refers to the preset weight coefficient corresponding to different first-level indicators A.

8. The method for evaluating the quality of postgraduate training based on an interval model according to claim 7, characterized in that: Horizontal comparative analysis involves comparing the Q values ​​of different evaluation objects. j1 And C(A) j1 By making comparisons, we can identify strengths and weaknesses, as detailed below: Two universities were selected as evaluation subjects, and the Q values ​​of University 1 and University 2 were compared. j1 Let them be Q1 and Q2 respectively, and then compare Q1 and Q2: If Q1 > Q2, then extract the comprehensive values ​​of the corresponding first-level indicator A for University 1 and University 2 respectively, and label them as C(A)1 and C(A)2 respectively; When C(A)1 > C(A)2, it means that university 1 is stronger than university 2 in the primary indicators corresponding to the quality of classified training, scientific research practice results, student development, or resource guarantee. Furthermore, if C(A)1 is higher than the corresponding preset first comprehensive threshold C(A) max This indicates that University 1 performs well overall in the corresponding primary indicator A, while C(A)2 is lower than the corresponding preset second comprehensive threshold C(A). min This indicates that University 2 has a problem with the corresponding primary indicator A.

9. The method for evaluating the quality of postgraduate training based on an interval model according to claim 8, characterized in that: Longitudinal comparative analysis compares the evaluation results of the same evaluation object at different time periods to analyze the development trend of its postgraduate training quality, as detailed below: First, select a university for Q corresponding to e time periods. j1 Then, according to the time sequence, the Q corresponding to each of the e time periods for each university is determined. j1 They are denoted as Q1, Q2, ... Q. e ; Next, calculate the difference between two adjacent time periods and denote it as QC. r Where r = 1, 2, ..., e-1, e-1 represents the number of differences between two adjacent time periods; Simultaneously calculate Q1, Q2, ..., Q e The average value is denoted as QP; Then, the difference QC between each two adjacent time periods is calculated. r The difference threshold QC between the two values ​​is... y Comparison: When QC r >QC y When the time interval is r+1, it indicates that the time interval is improved relative to the time interval r, and the period from the time interval r to the time interval r+1 is recorded as the improvement period; When QC r <-QC y When the time interval is r+1, it indicates that the time interval is decreasing relative to the time interval r, and the period from the time interval r to the time interval r+1 is recorded as the decreasing period; When -QC y ≤QC r ≤QC y When the time interval is r+1, it indicates that the time interval is relatively stable compared to the time interval r. Then, extract the number of rising and falling periods respectively, and denot them as TS and TJ respectively, and calculate the percentage of TS and TJ in e-1 respectively; At the same time, Q1 and Q e Comparison with QP: when >60%, Q e >Q1 and Q e >If QP is true, then it is determined that there is an overall improvement in e time periods; when >60%, Q e If both Q1 and QP are true, then it is determined that there is an overall decline in e time periods; If it is determined that there is no overall increase or decrease, it means that there is overall stability in the e time periods.

10. A postgraduate classification training quality evaluation system based on an interval model, the system being used to execute the postgraduate classification training quality evaluation method based on an interval model as described in any one of claims 1-9, characterized in that, The system includes: Model construction module: Selecting categorized training quality, scientific research achievements, student development, and resource support as primary indicators; establishing several secondary indicators under each primary indicator; for each secondary indicator, determining its minimum and maximum values ​​based on pre-collected historical data and relevant industry standards and inter-university comparative data; labeling each secondary indicator as i, where i = 1, 2, ..., n, where n is the total number of secondary indicators; subsequently constructing a corresponding single indicator interval [a] for each secondary indicator. i ,b i For each secondary indicator, experts in the field of postgraduate education are invited to assign weighted scores. Then, the average expert weighted scores for each secondary indicator i are calculated and recorded as the average expert score. Next, the sum of the average expert scores for all secondary indicators is calculated. The weight coefficient β for each secondary indicator i is obtained by dividing the average expert score by the sum of the average expert scores for all secondary indicators. i ; Data Acquisition and Processing Module: Multiple evaluation objects are selected from different universities, different majors, and different time periods within the same university. Each evaluation object is denoted as sample j, where j is the sequence number that distinguishes different samples. Various secondary indicators are collected from different channels and cleaned to handle missing and outlier values. Matching module: Each evaluation object is denoted as j1, where j1 includes the university, major, and time period of the same university. Then, the preprocessed data of each secondary indicator i is matched with its corresponding single indicator interval [a]. i ,b i Matching; Comprehensive evaluation module: The secondary indicator i corresponding to each evaluation object j1 is weighted; then, based on the weighting result, the comprehensive value of the primary indicator is determined; then, based on the comprehensive value of the primary indicator and combined with the preset weight coefficients corresponding to the primary indicator, the comprehensive evaluation result of the postgraduate classification training quality of evaluation object j1 is determined. Evaluation and Analysis Module: Performs horizontal and vertical comparative analysis on the comprehensive evaluation results of the calculated evaluation object j1 and the comprehensive values ​​of each primary indicator.

Images