Higher education quality evaluation based on neural network model

Abstract

With the constant goal of improving the quality of higher education, quality evaluation is a widely concerned problem, prompting this study to construct a higher education quality evaluation method based on a neural network model. First, an Attention Relevance Confidence Satisfaction (ARCS) model was constructed herein. This was done through two rounds of screening; the evaluation indexes of higher education quality were selected, and an evaluation index system was finally constructed. Then, the weight of each evaluation index was calculated using the constructed ARCS model. According to the 1–9 grading scale, an index scoring matrix of industry-education integration was established. Afterwards, the higher education quality evaluation score was obtained based on the neural network model, and the evaluation effect level was determined. The experimental results showed that the quality evaluation effect of the proposed method in the past five years showed an overall rising trend, even already reaching the top level. Moreover, the denoised experimental dataset was finally divided into a test dataset (28%) and an experimental dataset (72%), with the proposed method exhibiting a favorable effect.

Keywords

Neural network model higher education quality evaluation evaluation index system ARCS model

1. Introduction

With the continuous promotion of educational innovation, the objective evaluation of teaching quality has since become a vital link [1, 2]. From a relatively macro point of view, the evaluation methods of teaching quality involve teaching ability, teaching attitude, teaching content, teaching method, teaching effect, etc. [3]. These methods also contain several levels, measuring the teaching effect of different teaching factors. Hence, the combined action of these factors should be comprehensively analyzed to evaluate teaching quality [4, 5] accurately.

Given the skyrocketing of the country’s national economy, optimization and progress have also been continuously made in higher education. As an important part of China’s education, the quality evaluation of higher education is also geared towards being innovative. Among these, the role of higher education quality evaluation in higher education teaching is very important [6, 7]. Higher education quality evaluation ensures the improvement of the understanding of students’ comprehensive ability, including their comprehensive analysis skills, practical ability, innovative creativity, and capacity for self-studying. This kind of education quality evaluation, which combines theoretical knowledge and practical experience, significantly positively affects higher education work in China. Under the neural network model, teaching quality evaluation is a pivotal key link [8, 9, 10]. By constantly analyzing students’ learning effects, the shortcomings in the quality evaluation of higher education can be improved in real-time, thereby promoting the reform and innovation of higher education teaching and allowing for the proper implementation of higher education practice to achieve expected teaching outcomes.

Given the particularity and multi-dimensionality of practical teaching, most schools’ evaluation methods of teaching quality cannot meet current society’s educational needs. Therefore, modern technology and thought should be combined to develop a more scientific and reasonable evaluation method for practical teaching. Catering to the abovementioned background, the neural network model theory was analyzed herein, and a quality evaluation method of higher education was investigated. These were conducted with the expectation of providing a referable scheme for evaluating the teaching effect of higher education, ensuring the quality of practical teaching, promoting the sustainable development of practical teaching, and increasing the overall economic and social benefits of educational engineering.

Wang et al. [11] Following in-depth learning and fuzzy fault tree analysis, the quality evaluation of innovation and entrepreneurship education in colleges and universities. The quality of innovation and entrepreneurship education in colleges and universities is closely related to undergraduates’ learning degree of innovation and entrepreneurship knowledge and entrepreneurial motivation, making the effective quality assessment of education ultimately essential. The fault tree analysis has been commonly used to evaluate the quality of education. However, it has shortcomings, such as the dependence on fault data and the inability to deal with uncertainty.

Hence, the following steps were undertaken: First, a theoretical model based on the deep learning method was proposed to analyze the influencing factors of innovation and entrepreneurship education quality. Based on the traditional fault tree analysis, the fuzzy fault tree analysis was proposed to evaluate the reliability of classroom teaching in innovation and entrepreneurship education for college teachers and students. Finally, based on the top event of entrepreneurial teaching failure, a hyper-ellipsoidal model was implemented to limit the interval probability of basic events and describe the deviation of uncertain events.

Additionally, the accuracy of the model was verified using a questionnaire survey. Subsequently, the factors affecting innovation and entrepreneurship education quality were analyzed. In summary, the proposed method could effectively and quantitatively assess the quality of innovation and entrepreneurship education in colleges and universities, thus providing the basis for developing relevant improvement strategies to provide important technical support for improving the quality of innovation and entrepreneurship education.

Li et al. [12] applied a big data analysis model to the quality evaluation of talent cultivation in higher education. Given the unreasonable evaluation and unclear weight relationship of the assessment of education indexes, they proposed a big data analysis model to evaluate teaching evaluation indexes, providing more comprehensive scientific significance. Different systems in the index system were also taken as analysis objects, and the first-level weight relationship was normalized to make the weight distribution more reasonable. The teaching quality evaluation system thus became more sensible and scientific following the adoption of the big data analysis method.

In the same study, an index system of higher education background quality was designed and constructed, the weight relationship of different education indexes was analyzed using big data, and four main indexes were ultimately acquired. Afterwards, the weight relationship of the second-level indexes was analyzed and finally established. The results showed that the weight relationships of the four indexes were 0.3285, 0.1973, 0.2967, and 0.1755, respectively, and the evaluation model of educational quality was also given. Chakraborty et al. [13] proposed the influence of the quality of evaluation indexes on the variation of higher education labels. They stated that the evaluation results varied with different labels and university backgrounds under different circumstances, which would eventually harm students. In their survey, freshmen from Australian business schools were tested to understand multiple courses. The possible influences of evaluation guidelines and relevant document quality were explored, wherein it was found that the variation of labels was influenced by specific characteristics among evaluation criteria.

Zhu et al. [14] proposed evaluating college sports training quality using the Apriori algorithm. Using physical fitness training as the breakthrough point, the training quality data was evaluated following the Apriori algorithm, and factor mining was performed accordingly. The present practical teaching situation of physical training programs in colleges and universities in China was then deeply analyzed using the Apriori algorithm. Then, the advantages of sports training programs in college teaching were analyzed from the development of existing sports training programs and their functional characteristics. Meanwhile, suggestions on the current teaching model and the current situation of colleges and universities were forwarded to promote sports development in China. The study’s practical case analysis shows that physical training programs are feasible in college educational curriculums. To some extent, the spirit of sports training programs is similar to the concept of college physical education. There are also many sports training programs, which can also be used as rich resources for college physical education. Results reveal that the Apriori algorithm can effectively evaluate the quality of sports training. Male students’ mastery level of sports skills is generally higher than that of female students, and the difference is relatively significant. The algorithm proposed by the study effectively evaluates the data, which is important in supporting the decision to improve training quality.

Brown et al. [15] proposed the challenge of reflective practice in shaping the cognition of higher education academic managers on critical self-assessment reports. The study was limited to a single institution in the early stage with the application of CSER, with results showing the dependence on the reflective action of higher education academic management. The participants’ reflections were consistent with the reflection models in the references. The participants’ practice was on the wane and wax in the independent self-reflective practice. Moreover, the participants could independently complete self-evaluation in the programs, with collaborative communication and reflective abilities.

Su et al. [16] studied the quality evaluation system of the party-building work in colleges and universities in the new era. They pointed out that party-building work in colleges and universities plays an important role in various work in colleges and universities, which is a crucial guarantee in promoting colleges and universities to fulfil the mission of educating people and realizing long-term stable development. The current international political situation is getting increasingly severe. Like other grass-roots party organizations, a scientific and reasonable quantitative evaluation system must support party construction work in colleges and universities; otherwise, its quality and level cannot be objectively judged. Following the background and significance of the research on the quality evaluation system of the party building work in colleges and universities, the research results of this system obtained by domestic scholars were sorted out to provide some references for improving this system, improving the quality of the party building work in colleges and universities, and implement the paths and methods for the holistic development of colleges and universities.

Liu et al. [17] proposed applying the analytic hierarchy process (AHP) to construct an evaluation index system for the educational quality of nursing graduates. First, the indices of the evaluation index system for educational quality were determined using a literature review. Simultaneously, a Delphi questionnaire was designed, and 13 experts evaluated and scored the respondents invited to conduct two rounds of questionnaires. AHP and the percentage method then determined the weight of each factor, and the evaluation index system of postgraduate education quality for the nursing speciality was finally established. Following the weight calculated by AHP, the first-level indices were ranked in the following order: “input quality” (0.1273), “process quality” (0.3111), “output quality” (0.0846), and “development quality” (0.4770). In the second-level indexes, experts focused most on career development (0.3180). The top three of the third-level indexes were workplace (0.2385), personal expectations versus job opportunities (0.1272), and promotion opportunities (0.0795). The quality index system of nursing postgraduate education is shown to be scientific and reliable, and the weight allocation is also shown to be reasonable.

2. Methodology

2.1 ARCS model

Higher education teaching was deeply investigated in the current study using the ARCS model; the factors influencing students’ learning motivation were analyzed, their learning motivation was deeply aroused, and their active course learning ability was enhanced [18, 19, 20]. The ARCS model includes four parts, namely Attention, Relevance, Confidence, and Satisfaction, to motivate and encourage students to learn [21, 22]. The four factors in the ARCS model corresponded to the complete higher education teaching system process. They served as the four modules – definition, design, motivation, and evaluation – in the higher education teaching system. The specific structure of the ARCS model is shown in Fig. 1 above.

Figure 1.

Higher education teaching quality evaluation structure diagram of the ARCS model.

The definition module in the ARCS model mainly aimed to complete students’ conscious choice of higher education courses, cultivate their interest in exploring this lesson, and constantly direct their focus and attention on the teaching classroom and content [23, 24, 25]. The design module discussed the correlation between the courses learned by the students and students’ daily lives and increased the value of the course to the students after learning. The motivation module reflected how students established learning self-confidence through the learning environment and knowledge learning in the classroom [26, 27]. The evaluation module enhanced students; classroom teaching effect and improved their learning satisfaction through higher education teaching quality evaluation.

(1) Selection of first-stage evaluation indexes

In the first stage, the indices with relatively repeated functions were mainly screened out to reduce the repeatability of indexes. The process was as follows: First, relevant previous documents, materials, journals, and other materials were consulted, and all involved indices were collected as much as possible to form a comprehensive indicator system.

Step I: Assuming that $m_{i}$ represents the $i$ -th evaluation index, its mean value was calculated using the following formula:

$\displaystyle\bar{M}=\frac{\sum\limits_{i=1}^{n}{m_{i}}}{n},i=1,2,\ldots,n$ (1)

In Eq. (1), $n$ stands for the number of evaluations; $\sum\limits_{i=1}^{n}{m_{i}}$ represents the total sum of all the evaluation indexes for industry-education integration.

Step II: Each index was compared with the mean value $\bar{M}$ and processed using different standardization methods, which was as follows:

$\displaystyle\left\{{\begin{array}[]{l}\hat{M}_{i}=\frac{M_{i}}{\bar{M}},M_{i}% \leqslant\bar{M}\\ \hat{M}_{i}=\frac{\bar{M}}{M_{i}},M_{i}>\bar{M}\\ \end{array}}\right.$ (2)

In Eq. (2), $\hat{M}_{i}$ denotes the standardized $i$ -th evaluation index.

Step III: The correlation coefficient between the evaluation indexes was calculated using the following formula:

$\displaystyle\textit{Aij}=\frac{\sum\limits_{i,j=1}^{n}{(\hat{M}_{i}-\bar{M})}% (\hat{M}_{j}-\bar{M})}{\sqrt{\sum\limits_{i,j=1}^{n}{(\hat{M}_{ij}-\bar{M}_{j}% )^{2}}}}$ (3)

In Eq. (3), $\hat{M}_{j}$ represents the standardized $j$ -th evaluation index.

Step IV: The indices with repeated functions were screened out according to the following rules:

$\displaystyle\left\{{\begin{array}[]{l}A_{ij}>T,\to B\\ A_{ij}\leqslant T,\to C\\ \end{array}}\right.$ (4)

In Eq. (4), $T$ represents the set critical value, $B$ means deletion, and $C$ indicates reservation.

(2) Selection of second-stage evaluation indexes

In the second stage, representative indexes were screened out. In the remaining evaluation indexes, some indexes contributed little to the evaluation result, with their influences deemed as negligible. Hence, the final evaluation indexes were screened out in the second round, specifically done as follows:

Step I: A directed graph of the correlation between the remaining evaluation indexes was drawn, and the influencing degree between two indexes (marked as $j_{i}$ and $j_{j}$ ) was determined through the 1–3 scale method, thereby establishing an influence matrix $J$ , namely:

$\displaystyle J=\left[{{\begin{array}[]{*{20}c}{j_{11}}&{j_{1\mbox{2}}}&\ldots% &{j_{1m}}\\ {j_{21}}&{j_{21}}&\ldots&{j_{2m}}\\ \ldots&&&\\ {j_{m1}}&{j_{m1}}&\ldots&{j_{mm}}\\ \end{array}}}\right]$ (5)

Step II: Assuming that $B_{ij}$ indicates the influencing degree of the evaluation index $j_{i}$ on $j_{j}$ , the influencing degree of $j_{i}$ on $j_{j}$ and the degree to which $j_{i}$ was influenced by $j_{j}$ were calculated on basis of, which was as follows:

$\displaystyle B_{ij}=\frac{|{j_{i}-j_{j}}|}{\sum\limits_{j_{ij}\in A}^{m}{j_{% ij}}}\times\hat{B}_{ji}$ (6)

In Eq. (6), $\hat{B}_{ji}$ is the degree to which $j_{i}$ is influenced by $j_{j}$ .

Step III: The centrality $C_{i}$ of the index $j_{i}$ was calculated through the following formula.

$\displaystyle C_{i}=B_{ij}+\hat{B}_{ji}$ (7)

An evaluation index system was constructed by taking indexes satisfying $C_{i}\geqslant 1.0$ as the evaluation indexes for higher education quality.

2.2 Weight calculation of evaluation indexes for industry-education integration

The weight of each index, which is the proportion of the effect exerted by the index in the whole evaluation index system, was also calculated. Specifically, the index weights of industry-education integration were calculated using the designed ARCS model, which was as follows:

Step I: The selected indexes for industry-education integration were subjected to layered processing to establish a hierarchical evaluation index system.

Step II: $N$ evaluators were chosen to score each index according to the 1–9 grading scale, thereby establishing a scoring matrix $G$ of evaluation indexes for industry-education integration.

$\displaystyle G=(g_{pj})_{MN}=\left[{{\begin{array}[]{*{20}c}{g_{11}}&{g_{12}}% &\ldots&{g_{1N}}\\ {g_{21}}&{g_{22}}&\ldots&{g_{2N}}\\ \ldots&&&\\ {g_{M1}}&{g_{M2}}&\ldots&{g_{MN}}\\ \end{array}}}\right]$ (8)

In Eq. (8), $g_{pj}$ refers to the score given by the evaluator $p$ to the index $j$ , and $M$ denotes the number of indexes.

Step III: The objectivity coefficient of evaluators was calculated using the following formula:

$\displaystyle W=\frac{\bar{G}}{\sum\limits_{p,q=1}^{N}{G_{pq}}}$ (9)

In Eq. (9), $W$ denotes the objectivity coefficient of evaluators and $\bar{G}$ stands for the mean value of all net benefit factors.

Step IV: The index weight for industry-education integration was multiplied by the objectivity coefficient of evaluators to obtain the final evaluation index weight of higher education quality, which was as follows:

$\displaystyle Q=W\times H_{FD}$ (10)

In Eq. (10), $H_{FD}$ indicates the weight of each evaluation index for industry-education integration. Using the abovementioned process, the evaluation index weight calculation of higher education quality was thus completed.

2.3 Evaluation based on the neural network model

Neural network model is a deep neural network structure which has greater computing power than general neural networks. It is also more efficient and accurate in evaluating the quality of higher education. The neural network-based evaluation model is displayed in Fig. 2 above.

Figure 2.

Neural network-based evaluation model.

In the evaluation effect based on the neural network model, assuming that $b_{j}^{k}$ represents the bias term of the $j$ -th index at the $k$ -th layer, the evaluation index weight of higher education quality was used as the input for the convolution operation as follows:

$\displaystyle L_{j}^{k}=f\left(\sum\limits_{k=1}^{Z}\sum\limits_{j=1}^{M}W_{j}% \times Q_{j}^{k}\times v_{ik}+b_{j}^{k}\right)$ (11)

In Eq. (11), $Q_{j}^{k}$ indicates the convolution kernel function, $Z$ stands for the number of convolutional layers, $v_{ik}$ denotes the connection weight between the input-layer node $i$ and the convolution node at the $k$ -th layer, and $f$ indicates the activation function, here being the sigmoid function.

After the convolution layer processing, the obtained index convolution results entered the pooling layer and were sampled to reduce the dimensionality and the computational complexity. Finally, the Softmax classifier was used to calculate the score of each evaluation level, and the level corresponding to the maximum value was the evaluation result of higher education quality. Here, the evaluation result was divided into 5 levels, namely good, relatively good, ordinary, relatively poor, and poor.

Through the abovementioned process, the research on the higher education quality evaluation method based on the neural network model was then completed.

3. Results analysis and discussion

To better illustrate the effect and feasibility of the proposed higher education quality evaluation method based on the neural network model, an experimental link was constructed and the actual evaluation effect of the evaluation method was also analyzed.

3.1 Experimental environment

The comparison objects selected herein were the methods proposed in references [11, 12]. The selected evaluation index was also the degree of fitting with the actual evaluation result. A higher degree of fitting represented the higher evaluation accuracy of the evaluation method, which also indicated that it better facilitated the objective and accurate evaluation of higher education quality.

The experimental data was derived from the OpenCorporates database, which is one of the largest open databases. The test object selected here was a teacher from Zhejiang College of Security Technology. By investigating the teaching effect and teaching content, a judgment matrix was constructed using the teaching quality evaluation method, and the corresponding index weight was thus calculated. The online teaching quality evaluation data for teachers in the first semester of 2022 was randomly selected as the standard to verify the accuracy of the evaluation method. To make the evaluation results more accurate and objective, ten experienced teachers in the school were consulted, and a judgment matrix about their teaching quality s was obtained, as seen in Table 1 above.

Table 1
Specific parameters of teaching quality judgment matrix

A	B1	B2	B3	B4	$\gamma_{\max}$	R
B1	1	1	3	4	4.02	0.03
B2	1	1	3	4
B3	1/3	1/3	1	2
B4	1/4	1/4	1/2	1

According to the teaching quality judgment matrix constructed above, the judgment weight was calculated to obtain a specific evaluation score, and the accuracy of the three evaluation methods was compared using the difference between the evaluation score obtained by the judgment method and the historical teaching quality evaluation data.

3.2 Evaluation index system

The relevant parameters of the neural network model were set as follows: 1) the size of the convolution kernel was 20 $\times$ 20; 2) the convolution step size was 8; 3) the pooling step size was 6; the number of neurons was 15, 25, 35, 45, and 55; 4) the number of training steps was 200; 5) the learning rate was 1.15; and 6) the training accuracy was 0.0001. The methods proposed in references [11, 12] were compared, and the evaluation results are displayed in Fig. 3 below.

Figure 3 shows that the quality evaluation effect of this school in the recent 5 years was evaluated using three different methods, therein showing an overall year-on-year growth trend. The evaluation effect achieved therein by the method proposed in this research reached the top level, while that by the other two methods was favorable with certain room for improvement.

3.3 Experimental dataset

The related teaching quality evaluation data of college education class acquired by the neural network model were improved using CADD data, and both types of data served as two experimental datasets, with their specific information seen in Table 2 below.

Table 2
Comparative information of two experimental datasets

Item	Name of experimental dataset
	Dataset I	Dataset II
Data size	50 GB	100 GB
Data pieces	235421 pieces	212154 pieces
Course type	Traditional course $+$ MOOC	Traditional course
Teaching environment	Teaching classroom $+$ network learning environment	Teaching classroom
Data span	4 years	3 years
Data correlation	0.94	0.83

The teaching quality evaluation data contained in each dataset included pre-class data, in-class data, after-class data, along with the activity data of teachers, the environment, and of students. For the evaluated experimental dataset, data denoising was performed using the neural network model, with the information on the denoised experimental dataset listed in Table 3 below.

Table 3

Information of the denoised experimental dataset

Item	Name of experimental dataset
	Dataset I	Dataset II
Data correlation	0.97	0.98
Average denoising accuracy of events	0.9453	0.9745
Evaluation result	28%	72%

Figure 3.

Evaluation results of higher education quality under the three methods.

After the denoising processing, the performance of teaching quality evaluation of the design method was tested using the denoised experimental data set. The denoised experimental dataset was then randomly divided into training data and test data. The final division result was a test dataset (28%) and an experimental dataset (72%). The designed higher education quality evaluation method based on the neural network model then ultimately harvested a good effect.

3.4 Discussion

Given the evident diversity of factors affecting the evaluation of higher education quality, to ensure that the subsequent evaluation result can reflect the actual teaching quality to the greatest extent, talent cultivation plays a crucial role in the long-term national development. Hence, educational construction and perfection have long-term priorities and national concerns. In researching the evaluation of higher education quality, the ARCS model-based quality evaluation structure of higher education teaching was applied, the weight of each evaluation index was calculated, the level of the evaluation effect was determined, and the innovative, diversified, and comprehensive higher education quality evaluation was realized from the technical level, which provides much enlightening significance to the research on teaching quality evaluation and lays a good foundation in the follow-up of accurate and scientific quality evaluation. Moreover, the neural network model was also applied to the design method to mine and process the teaching quality evaluation data, thereby providing data support for teaching quality evaluation. Against this background, compared with the methods mentioned in relevant references, the method proposed herein is more capable of accelerating the adjustment and improvement of the development scheme for higher education and shows great realistic significance for clarifying the evaluation effect of higher education quality. It is, therefore, expected to improve the evaluation efficiency of higher education quality given the powerful ability of the neural network model.

4. Conclusion

This study forwarded a quality evaluation method of higher education based on the neural network model. First, an ARCS model was constructed, the evaluation indices for higher education quality were selected through two rounds of screening, and an evaluation index system was established. Afterwards, the weight of each evaluation index was calculated. Next, the evaluation score of higher education quality was obtained based on the neural network model, and the level of the evaluation effect was determined, thereby completing the research on the quality evaluation method of higher education. Finally, the following conclusions were drawn: the evaluation effect achieved by the proposed method has already reached the top level, and the quality evaluation effect shows an overall year-on-year growth trend. The final division result is thus a test dataset (28%) and an experimental dataset (72%), and the designed neural network model-based higher education quality evaluation method displays an excellent effect.

Footnotes

Acknowledgments

This study was supported by the Department of China Association of Higher Education (Program No. 22GDZY0211).

References

Bueche

Jensen

Martin

Riddle

Stote

. Distance education in nutrition and dietetics education over the last 30 years: A narrative review. Journal of the Academy of Nutrition and Dietetics. 2022; 23(11): 11006.1-11006.14.

Handoko

. The role of service excellence training for government employees to improve customer satisfaction: Public service implementation in tax service office in jakarta. Technium Social Sciences Journal.2021; 16: 1.

Gupta

Viswesh

Cone

Unni

. Qualitative analysis of the impact of changes to the student evaluation of teaching process. American Journal of Pharmaceutical Education.2020; 84(1): 7110.

Shen

. Quality evaluation of entrepreneurship education in Chinese medical colleges-from the perspective of student cognition. Frontiers in Psychology.2020; 11: 1093.

Domenici

. A course of history of chemistry and chemical education completely delivered in distance education mode during epidemic COVID-19. Journal of Chemical Education.2020; 97(9): 2905-2908.

Madadi

Bertiche

Escalera

. SMPLR: Deep learning based -MPL reverse for 3D human pose and shape recovery. Pattern Recognition.2020; 106: 107472.

Nie

. Fuzzy evaluation model of the teaching quality in colleges and universities based on analytic hierarchy process. Basic & Clinical Pharmacology & Toxicology.2020; 127(1): 189-189.

. valuation and systematic construction of physical education quality based on integration. Journal of Tianjin University of Sport.2021; 36(5): 505-511.

. Research on evaluation system of high-quality development of industry characteristics universities. Chinese Geological Education.2021; 30(2): 6-9.

10.

Liang

Huang

Long

Zhang

. Deep Reinforcement Learning for Resource Protection and Real-time Detection in IoT Environment. IEEE Internet of Things Journal.2020; 7(7): 6392-6401.

11.

Wang

Zheng

Zhang

Qian

Zhang

Zou

. Exploring quality evaluation of innovation and entrepreneurship education in higher institutions using deep learning approach and fuzzy fault tree analysis. Frontiers in Psychology.2022; 12(1): 5981-5992.

12.

Wang

Bian

. Application of the big data analysis model in higher education talent training quality evaluation. Complexity.2021; (31): 1-9.

13.

Chakraborty

Dann

Mandal

Dann

Paul

Hafeez-Baig

. Effects of rubric quality on marker variation in higher education. Studies In Educational Evaluation.2021; 70(9): 100997.

14.

Zhu

, Wang Yu

Wang

. Quality evaluation of college physical training considering apriori algorithm. Mathematical Problems in Engineering.2022; (8): 1-8.

15.

Brown

. The challenges of reflective practice in shaping higher education academic managers perceptions of critical self-evaluation reports in quality enhancement. Allied Business Academies.2021; 20: 1-18.

16.

. The background significance and results combing of the research on the quality evaluation system of the party building work in colleges and universities in the new era. Journal of Higher Education Research.2(4): 178-185.

17.

Wei

Liu

Kang

Luo

Zhang

. Evaluation index system of education quality for nursing professional degree postgraduate using the analytic hierarchy process. Medicine.2021; 100(47): e27771.

18.

Shlobin

Radwanski

Sandhu

MRS

Rosseau

Dahdaleh

. Increasing equity in medical student neurosurgery education through distance learning. World Neurosurgery.2022; 32(11): 63-187, 196,

19.

. Corpus-based English subtitle translation quality evaluation system design. Automation & Instrumentation.2020; 20(02): 81-84.

20.

Zheng

. Build online education quality evaluation system based on “four evaluations”. Educator.2021; (15): 36-37.

21.

Yuan

Zhang

Wang

. Design and empirical study of IPA-based evaluation model of public sports service quality. Journal of Chengdu Sport University.2020; 46(1): 60-66.

22.

Song

Zhao

. Asynchronous transmission and update simulation of remote real-time courseware data. Computer Simulation.2020; 37(11): 199-202, 226.

23.

Liu

Bai

Zhang

Jiang

. Empirical Analysis on the Quality Evaluation of Master’s Thesis in Colleges and Universities of Traditional Chinese Medicine: Taking Henan University of Traditional Chinese Medicine as an Example. Medical Education Research and Practice.2020; 28(4): 615-620.

24.

Gao

. Deep learning model based on improved point cloud transformer. Journal of North University of Chinaï¼ˆNatural Science Edition).2021; 42(6): 515-523.

25.

Qian

Huang

. Research on the quality evaluation of Chinese preschool education based on user satisfaction. Education Exploration. 2020; (3): 26-30.

26.

Shen

Dai

. Binary tree multi-classification algorithm based on weighted fuzzy membership. Application Research of Computers.2020; 37(11): 3281-3286.

27.

Mayo

Wong

. Starting off on the right foot: Providing timely feedback to learners in quality improvement education. BMJ Quality & Safety.2021; 31(4): 263-266.

Higher education quality evaluation based on neural network model

Abstract

Keywords

1. Introduction

2. Methodology

2.1 ARCS model

(1) Selection of first-stage evaluation indexes

(2) Selection of second-stage evaluation indexes

3.1 Experimental environment

Table 1 Specific parameters of teaching quality judgment matrix

3.3 Experimental dataset

Table 2 Comparative information of two experimental datasets

4. Conclusion

Footnotes

Acknowledgments

References

Table 1
Specific parameters of teaching quality judgment matrix

Table 2
Comparative information of two experimental datasets