RESEARCH ON THE REALIZATION
MECHANISM AND EVALUATION SYSTEM
OF HIGH-QUALITY UNDERGRADUATE
EDUCATION IN PRIVATE UNIVERSITIES
BASED ON DEEP LEARNING
Jun Wang*
Guangdong Institute of Science and Technology, Dongguan, Guangdong, 523808,
China
wjshgz2022@126.com
Reception: 22/02/2023 Acceptance: 17/04/2023 Publication: 16/05/2023
Suggested citation:
Wang, J. (2023). Research on the Realization Mechanism and Evaluation
System of High-Quality Undergraduate Education in Private Universities
Based on Deep Learning. 3C TIC. Cuadernos de desarrollo aplicados a las
TIC, 12(2), 97-115. https://doi.org/10.17993/3ctic.2023.122.97-115
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
97
ABSTRACT
Due to the new development stage, it is especially important to improve the education
quality of private undergraduate universities. As a result, it is a new hot issue for the
construction of a mechanism and assessment system for the quality improvement of
private undergraduate education. In this paper, after analyzing and researching the
quality of undergraduate education in present-day universities, the mechanism of
deep learning is applied to the establishment of the assessment system. Finally, 1082
samples collected from the data center platform of a private university are analyzed as
the research object. From the results, the final size of the combined weights of the
seven evaluation items constituting the assessment system differed basically little.
They were 12.81%, 15.78%, 15.28%, 14.38%, 12.83%, 12.81%, 15.01%, and
13.27%, respectively. In the comparison of this paper's method with FAHP+TOPSIS
combined evaluation, euclidean map method, and genetic algorithm assignment, the
difference between the seven weight values of the euclidean map method is larger,
5.56%. The evaluation times of the four methods were 41 s, 38 s, 47 s, and 118 s.
Compared with the other three methods, the genetic algorithm assignment took the
most time.
KEYWORDS
Deep learning; private universities; high-quality undergraduate education; realization
mechanism; assessment system
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
98
INDEX
ABSTRACT
KEYWORDS
1. INTRODUCTION
2. THEORETICAL BASIS OF THE BLENDED EDUCATION MODEL BASED ON DEEP
LEARNING
2.1. Deep Learning
2.2. Blended teaching
2.3. DELC model
3. MECHANISM FOR REALIZING HIGH-QUALITY UNDERGRADUATE EDUCATION
IN PRIVATE UNIVERSITIES BASED ON DEEP LEARNING
3.1. Pre-course shallow learning
3.2. Knowledge framework construction
3.3. Knowledge depth processing
4. CONSTRUCTION OF ASSESSMENT SYSTEM FOR PRIVATE UNDERGRADUATE
EDUCATION
4.1. Evaluation Model
4.2. Data Acquisition
4.3. Data pre-processing
4.4. Variational CRITIC Assignment Method
4.5. Improving the entropy weight method
4.6. Compound weights
5. EXPERIMENT AND ANALYSIS
5.1. Credibility Analysis
5.2. Comparison of algorithms
6. DISCUSSION
7. CONCLUSION
REFERENCES
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
99
1. INTRODUCTION
Stepping into the new century, higher education in China has ushered in a new
period [1]. With the continuous improvement of education methods, higher education
has moved from the previous high growth rate to a new stage of high-quality
development [2-4]. The improvement of teaching quality is not only for public
undergraduate institutions. For private undergraduate institutions, seizing the
opportunity to improve teaching quality is a decisive factor for their further
development in the future. Most of the private undergraduate institutions are formed
by upgrading local higher education institutions as the basis. They highlight their
regional characteristics in terms of schooling characteristics while taking the needs of
talents in the region as their own teaching orientation [5-7]. In the context of China's
vigorous development of high-quality education, there are two aspects that need to be
focused on. The first is how to seize the policies favorable to the development of
private universities and add reliance to the development. The second is how to
increase the kinetic energy within the school to comply with the development call and
form a unified consensus. Creating private higher education institutions with certain
influence as well as teaching strength should be placed at the top of the task list of the
management of these universities [8].
In improving the quality of college education, numerous scholars have conducted
research at various levels. Zeng Y [9] studied the influence of teachers' teaching
ability and students' level on the effectiveness of college physical education. In order
to further improve the accuracy of college physical education evaluation, he analyzed
the course teaching quality evaluation system of college physical education. After
analyzing the data mining techniques and the applicability of Hidden Markov in the
evaluation of the teaching quality of college physical education, the corresponding
mathematical model was proposed. The mathematical model was also validated by a
series of experiments. heng Q [10] explored the role of 5G technology in reforming the
quality of classroom teaching in colleges and universities under the accelerated
development of information technology. In his study, he built the teaching quality
system based on the B/S architecture model, using the SSM framework and MYSQL
database development. He believes that this can improve the school's teaching
quality evaluation system as a whole and make the system more objective and fair.
Hong W [11] believes that computer teaching, as an important part of college student's
education, can better promote students' all-around development for the improvement
of students' computer ability application. However, there are more problems in
computer education nowadays, and the teaching effect is not ideal. In his study, he
focuses on analyzing the problems of computer teaching in colleges and universities
in the environment of big data. Xu Z [12] believes that the construction of a practical
teaching quality system and its optimization are especially important in improving the
teaching quality of applied undergraduate institutions. In his study, the DEA model is
used to construct the evaluation indexes of practical teaching in applied
undergraduate institutions. Xue K [13] focused on analyzing the problems of teacher
quality evaluation in higher education institutions and proposed a correct view of
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
100
quality evaluation and an evaluation system with vocational and technical
characteristics. He believed that a suitable teaching quality evaluation system should
be reasonable and relevant. Only under the guidance of scientific evaluation methods
can a correct and developmental teaching system be established. anzhao M [14]
believes that the quality of practical teaching is the core competitiveness of private
applied new undergraduate institutions. He believes that these types of institutions
should focus on cultivating application-oriented talents. In order to achieve this goal, it
is important to clarify the content of monitoring the quality of practical teaching. In his
study, the focus is on the establishment of a practice teaching quality system. The
sustainable and supervisable teaching model is explored.
At a time when high-quality development is advocated, undergraduate education in
private universities should also join the flood of development. Apply the national
policies as well as their own resource advantages to their own development [15, 16].
After examining their own strengths and weaknesses, they should integrate them into
the new demands of development. It is also not negligible to find one's own position in
the development. With these needs, an appropriate teaching evaluation assessment
system is necessary to exist. Based on this evaluation system, it can play a role in
monitoring and improving the management level of the school, the quality of teachers'
teaching, and the development of students' abilities. It has a significant role in
significantly improving the level of education and competitiveness of the school. This
is the reason why this paper conducts a study on the mechanism and evaluation
system for achieving high-quality undergraduate education in private universities.
2. THEORETICAL BASIS OF THE BLENDED
EDUCATION MODEL BASED ON DEEP LEARNING
2.1. DEEP LEARNING
In the 1980s, Russell Ackoff's four-layer wisdom data framework bridged the gap
between "technology" and "wisdom" for intelligent education in the information age.
The framework demonstrates the evolutionary path of human intelligence from
information to knowledge and from knowledge to intelligence. In this process, people
build a knowledge network in their minds through active understanding of memory,
establish interrelationships among things and between things, understand, connect,
transfer, apply and create knowledge, and develop intelligence and wisdom. In the
process of knowledge development, deep learning that focuses on knowledge
internalization and transfer is crucial.
Deep learning theory has some similarities with deep learning of machines in the
field of artificial intelligence [17, 18]. The idea of machine learning is to build a
multilayer artificial neural network. During the training of the network, feedforward
operations are performed layer by layer on the input data, the final result of the output
operation is compared with the target, and the error is returned to each neuron to
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
101
adjust the connection parameters of each neuron. Multilayer neural network
connections can accurately represent the complex nonlinear mapping relationships
between input and output data [19-21]. It is similar to the internalization and
construction of knowledge in the human learning process. Among them, the functions
of evaluation and feedback are very important. Artificial neural networks have self-
identification and self-adaptive capabilities, and the performance of artificial
intelligence in many fields has now surpassed human capabilities. In recent years,
machine migration learning has become an important development direction, showing
great commercial value. Compared with intelligent machines, the network structure of
the human brain is more complex and profound, so it is equipped with better
conditions for deep learning [22-24]. The purpose of human deep learning is also the
transfer of knowledge. Therefore, this brings a profound insight into the realization of
current high-quality undergraduate education in colleges and universities: deep
learning is an important way to achieve the goal of current high-quality undergraduate
education in colleges and universities, and it is a bridge between information
technology and high-quality undergraduate education in colleges and universities.
2.2. BLENDED TEACHING
In the late 20th century, a blended learning theory was proposed in the West, and
today's blended teaching is derived from the blended learning theory. 2003, Professor
He of Beijing Normal University first proposed the concept of blended teaching in
China. He believes that blended teaching combines the advantages of traditional
teaching methods and online teaching. It not only plays a leading role in guiding,
facilitating and monitoring the teaching process and fully reflects the initiative of
teachers, but also shows the main features of students' enthusiasm and creativity in
the learning process. Blended teaching is not a superposition of traditional offline and
online teaching, but a rearrangement of teaching objectives, a reconfiguration of the
knowledge system and a curriculum design using appropriate teaching methods
based on a full analysis of their benefits in order to achieve the best teaching effect.
2.3. DELC MODEL
The DELC model is a deep learning route model proposed by American scientists
Eric Jensen and LeAnn Nickelsen in their book "7 Effective Strategies for Deep
Learning". The DELC model describes the entire design process of deep learning,
from the initial "design of standards and curriculum" to the final "evaluation of student
learning". It is a complete and clear pathway from lesson planning to implementation
and evaluation, leading students to deeper learning step by step [25]. Blended
teaching based on deep learning aims to make full use of the advantages of
information and teaching technology to realize the organic combination of deep
learning and blended teaching mode, to solve the problem of superficial learning to a
certain extent, and to promote the realization of deep learning.
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
102
Figure 1. "Deeper Learning Cycle" (Deeper Learning Cycle)
3. MECHANISM FOR REALIZING HIGH-QUALITY
UNDERGRADUATE EDUCATION IN PRIVATE
UNIVERSITIES BASED ON DEEP LEARNING
3.1. PRE-COURSE SHALLOW LEARNING
In the pre-course shallow learning stage, according to the first stage of the DELC
deep learning route "Designing Standards and Curriculum", i.e., the formulation of
curriculum objectives, combining Bloom's classification theory of teaching objectives
with the three-dimensional objectives of classroom teaching, the teaching objectives
can be divided into knowledge objectives, skills objectives and emotional objectives.
In deep learning, the goal of knowledge is not only the mastery of knowledge, but also
the deep understanding of knowledge. Students actively construct their own
knowledge system by activating their prior knowledge. Skill goals emphasize students'
ability to use their knowledge flexibly, including the ability to learn independently,
communicate collaboratively, and solve problems [26, 27]. Emotional goals, on the
other hand, focus on students' emotional experiences throughout the learning
process. Based on the mastery of knowledge and improvement of skills, students love
learning and enjoy the whole learning process. Teachers can design lessons guided
by clear teaching objectives. According to the characteristics of blended teaching,
teachers should break the traditional textbook system, reorganize more meaningful
teaching units, and collect, develop and adapt curriculum resources according to
students' career orientation, curriculum features and student characteristics. Before
the whole course starts, teachers can upload the syllabus, teaching plan and other
relevant teaching materials on the teaching platform. Students can have a certain
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
103
understanding and their own view of the classroom-related content by previewing the
materials [28].
3.2. KNOWLEDGE FRAMEWORK CONSTRUCTION
When first building a knowledge framework, teachers should first complete the third
step of the DELC Deep Learning Pathway, "Creating a Positive Learning Culture" and
the fourth step, "Preparing and Activating Prior Knowledge. To create a positive
learning culture, the first step is to establish a harmonious and reliable relationship
between the teacher and the students. A good relationship between teachers and
students will be further developed if teachers give students affirmation,
encouragement and proper guidance; at the same time, students should develop
harmonious and reliable interpersonal relationships between students through
communication and collaboration. Thus, a positive learning culture can provide an
emotional foundation for mastering new knowledge [29, 30]. Second, we should
consolidate the basic knowledge base, i.e., teachers should "prepare and activate
prior knowledge". In the process of pre-assessment, teachers learn about students'
prior knowledge in various ways. During this phase, teachers can activate students'
prior knowledge through tests, questions, surveys, discussions, and other methods.
Since most students lack prior knowledge, teachers should add relevant knowledge to
the curriculum so that students can connect old and new knowledge. This makes it
easier for students to learn new knowledge so that they can deepen their
understanding of it.
3.3. KNOWLEDGE DEPTH PROCESSING
In the early stages of knowledge construction, students' learning problems are
often complex, both the same and different. Some problems can be solved through
iterative research, while others require students to collect and analyze in different
ways; other problems can be solved through discussion among students or comments
by the teacher. No matter what the problem is and how it is solved, it lays a good
foundation for the third stage of deep learning, which is "deep processing of
knowledge". In the process of deepening knowledge, teachers should first answer the
most common questions that students ask during the learning process. For important
points, teachers should help students to reinforce their impressions of these points.
For some difficult problems and tasks, teachers should further organize and guide
students, such as ways to find information and problem-solving skills, to combine with
various practical problems encountered in daily life, to develop logic, rigor and
integrity of thinking in the process of problem solving, and to promote deeper
processing of knowledge. After solving problems, teachers should summarize key
knowledge and difficult knowledge to avoid the dispersion of knowledge caused by
online learning, help students form a systematic knowledge system, reorganize the
problem-solving process, and let students gain deeper learning experience and
achievement experience.
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
104
4. CONSTRUCTION OF ASSESSMENT SYSTEM FOR
PRIVATE UNDERGRADUATE EDUCATION
4.1. EVALUATION MODEL
The traditional teaching evaluation model is mainly a quantitative superposition of
individual indicators, which cannot reflect the hierarchical and systematic
characteristics of indicators. In order to comprehensively evaluate the teaching level
and educational success of high-quality undergraduate education in colleges and
universities, help schools and teachers improve their teaching and identify problems,
a systematic teaching evaluation model is established for this purpose according to
the evaluation process and work requirements (see Figure 2). The evaluation model
obtains evaluation index items from the training plan and course objectives, and then
sets the values of each evaluation index by investigating and comparing them, and
finally summarizes and analyzes the teaching effectiveness according to this
evaluation system.
Figure 2. Diagram of teaching assessment processing model
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
105
4.2. DATA ACQUISITION
Evaluation data can be obtained from the school's stored big data, and
subsequently obtained in the infrastructure, business platform, teaching system and
digital resource library according to the teaching evaluation objectives, with reference
to the requirements of the index system of school route, teaching conditions,
professional and curriculum construction, teaching management, academic style
management and quality cultivation. With the help of third-party big data management
platform tools, we collect and process various data, and evaluate the platform-
generated teaching evaluation data with the content required by the evaluation index
system in correspondence.
4.3. DATA PRE-PROCESSING
Let the level 1 index system set , dimensions be the
weight of the th rated index at level 3, , and the appraisal
sub-item score
, then the
evaluation value of a level is calculated as:
(1)
The initial data are de-quantified and then standardized. According to the algorithm
model of the undergraduate education assessment index, the mean of the th item
, standard deviation , and the mean is
normalized to:
(2)
Also standardized scores of:
(3)
Thus each indicator within will have ratings , and so
the rating matrix is obtained:
I={I1,⋯,I7}
i,j,k,wijk
k
w
ijk =
{
wij1,⋯,wijK
}
s
ijk =
{
s1
ijk,⋯,sL
ijk
}
,j∈J,l∈L,J,K,L∈Z
+
x′
ijk =
K
∑
k=1
wijk
L
∑
l=1
sl
ijk
i
¯
x
′
i=
1
J
J
∑
j=1
x′
i
j
S
i=1
J−1
J
∑
j=1
(x′
ij −¯x′
i)2
x*
ij =
x′
ij
−S
i
max (
x
′
i)−min (
x
′
i)
x
ij =
x*
ij
−¯x
i
Si
Ii
m
(m∈[1,M], M∈Z+)
Xi
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
106
(4)
Also indicator weight then the
indicator score of :
(5)
4.4. VARIATIONAL CRITIC ASSIGNMENT METHOD
The CRITIC assignment method can objectively express the difference or
correlation of evaluations of indicator , and its physical meaning is that the larger
the standard deviation is, the greater the role of the indicator; the correlation
coefficient is used to express the degree of conflict between indicators, and if the
correlation coefficient is higher, it indicates that the conflict between indicators is
smaller. In order to speed up the big data computation, the correlation strength is
calculated using the equivalent Pearson coefficient:
(6)
Obtain the evaluation correlation matrix , where
, then there is indicator conflictive ness :
(7)
(8)
where the coefficient of variation reflects the average variability and is the
adjustment factor to fit the integrated weight interval, generally , so that the
weight is obtained:
X
i=
x
1
i1x
2
i1Lx
M
i1
x1
i2x2
i2Lx1
i2
M M L M
x1
ij x2
ij LxM
iJ
w
ij =
K
∑
k=1
wijk,wi=
J
∑
j=1
wij,W=
J
∑
j=1
wij =
1
S
i=
∑M
m
∑J
j=1
xm
ij
w
ij
M∑J
j=1
wij
m
Ii
S
ρμ
i(j,q) = 1
M−1
M
∑
m=1 (xm
ij
−¯xm
ij
Sm
ij
)M
∑
m=1 (xm
iq
−¯xm
iq
Sm
iq
)
m
P
i
=(
ρ
i
)j×q
j,q∈[1,j], j∈Z+
Ci
C
i=δi
J−1
J
∑
j=1
1−
j
∑
q=1
ρi(j,q)
δ
i=r
x
i
−¯x
¯x
M
∑
m=1
(xm
i/ ¯xm)
δij
r
r= 2.5
Ii
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
107
(9)
4.5. IMPROVING THE ENTROPY WEIGHT METHOD
The entropy value can reflect the amount of information carried by an indicator, and
the larger it is, the higher the variability among related indicators, and the greater the
weight and role of an indicator in the whole evaluation system. Thus the process of
calculating the weight of the term:
(10)
Calculate the entropy value :
(11)
If , then , followed by calculating the variance:
(12)
Obtain the entropy weights:
(13)
In the entropy value, expert assignment values are introduced to
improve the overall entropy weights:
(14)
4.6. COMPOUND WEIGHTS
In order to achieve a comprehensive and objective reflection of the teaching level
and the subjective judgments of the parties in the evaluation, the closest combined
weight is calculated using the least squares method. is
known, and the shortest distance to is calculated, then there exists:
w
′
i=Ci/
I
∑
i=1
C
i
Ii
P
ij =
M
∑
m=1
xm
ij /
J
∑
j=1
M
∑
m=1
x
m
ij
ei(0≤ei≤1)
e
i=−
J
∑
j=1
Pij ln pij /ln
J
pij = 0
pijlnpij = 0
di= 1 −ei
w
ei =di/
I
∑
i=1
d
i
wS
q(q∈[1,I])
w
′ ′
i=
we
i
,i=q
we
i+we
i(we
q−ws
q)
∑q−1
i=1 we
i+∑I
i=q+1 we
i
,i≠
q
wi
w′
i,w′ ′
i,i∈[1,I], I∈Z+
w′
i,w′ ′
i
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
108
(15)
When the function is continuously differentiable, to simplify the calculation of
large data weights, the inverse of is directly solved to obtain the optimal
combination weights as:
(16)
5. EXPERIMENT AND ANALYSIS
5.1. CREDIBILITY ANALYSIS
Of the 1082 samples collected from the data center platform of a private university,
credibility analysis was first done and Cronbash's factor was introduced to test the
level 1 teaching evaluation credibility of the samples:
(17)
where the number of samples , is the variance of the score of the th
sample, and is the total variance to obtain Figure 3. The specific data values are
shown in Table 1. The variable weights and expected weights of the level 1
instructional evaluations of the different factors in the sample can be seen in Figure
3(a). The final combined weights obtained are shown in Figure 3(b), which consists of
a combination of variable weights and expected weights. The specific results are
calculated by equations (15)~(16). It is easy to see that the final size of the combined
weights of the seven evaluation items differed basically little, 12.81%, 15.78%,
15.28%, 14.38%, 12.83%, 12.81%, 15.01%, and 13.27%, respectively. This reflects
the fact that each item plays a similar role in the quality rating index. In terms of
individual evaluation items, the difference between the variable and expected weights
of instructional management is relatively significant compared to the other six items,
reaching 0.90%. The reason for this situation is the excessive uncertainty in
instructional management. This uncertainty makes it necessary to introduce correction
factors appropriately for improvement in the subsequent evaluation of the assessment
system. Meanwhile, the distribution of weights in Figure 3 (a) is not significantly
different from the requirements in the book "Indicator System" published by the
Ministry of Education, which proves the reliability of the algorithm proposed in this
paper. And in Figure 3 (b), the reliability of the 7 items of Level 1 teaching evaluation
are 0.89, 0.88, 0.90, 0.91, 0.87, 0.84, 0.89. Usually, α in indicates unreliable indicates
reliable, and greater than 0.9 indicates very reliable. From Figure 3, it can be seen
min
(
f
(
wi
))
=∑1
i=1
(
wi−w′
i
)2
+
(
wi−w′ ′
i
)2
s.t. ∑I
i
=1
wi= 1
f(wi)
f′ (wi)=0
wi
wi=(w′
i+w′ ′
i)/2
a
=n
n−1
(
1−
∑n
i=1
S2
i
S2
)
n= 1082
S2
i
i
S2
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
109
that the calculation results of the sample data can meet the accuracy requirements of
the index, thus ensuring the reliability of the collection of large sample data.
Figure 3. Weights of rating indicators and credibility of teaching evaluation of high-quality
undergraduate education in colleges and universities
Table 1. Data values for each comparison of credibility of level 1 teaching evaluation
5.2. COMPARISON OF ALGORITHMS
In this paper, we also selected the more mainstream weight assessment algorithms
for comparison, respectively, FAHP+TOPSIS combined assessment, superior order
diagram method, and genetic algorithm assignment, using uniform samples and index
coefficients, to obtain the weight of each index, and the results are shown in Figure 4.
The specific values of the comparison items of the four algorithms are shown in Table
2.
Evaluation Item
Contrast Item
School
Course
(%)
Teachers
Troop
(%)
Teaching
Conditions
(%)
Major and
Curriculum
Construction
(%)
Teaching
Management
(%)
Study Style
Construction
(%)
Quality
Education
(%)
Variable Weight 12.37 15.24 15.79 14.39 13.73 15.16 13.46
Expert Weight 13.25 16.33 14.81 14.38 11.94 14.87 13.08
Combination
Weight 12.81 15.78 15.28 14.38 12.83 15.01 13.27
Credibility 0.89 0.88 0.90 0.91 0.87 0.84 0.89
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
110
Figure 4. Comparison chart of various weighting algorithms
Table 2. Specific values of each comparison term for the four algorithms
The weighting of the indicator weights of the seven Level 1 teaching evaluations
under different methods is reflected in Figure 4. Among them, the seven weight values
of the Euclidean diagram method varied widely. Teaching tools have the largest
weight in the Euclidean method with 17.34%, and teaching style construction has the
lowest weight with 11.78%. The difference between the two was 5.56%. For the
assessment system of FAHP+TOPSIS and the genetic algorithm, the index weights of
each item are basically the same as the method proposed in this paper, and the
Evaluation
Item
Contrast
Item
School
Course
Weight
(%)
Teachers
Troop Weight
(%)
Teaching
Conditions
Weight
(%)
Major and
Curriculum
Construction
Weight (%)
Teaching
Management
Weight
(%)
Study Style
Construction
Weight
(%)
Quality
Education
Weight
(%)
Evaluation
Algorithm
Time
(S)
Method
Proposed In
This Paper 12.81 15.78 15.28 14.38 12.83 15.01 13.88 41
Optimal
Sequence
Diagram
Method
14.55 17.34 13.92 15.71 13.74 11.78 12.96 38
Genetic
Algorithm
Evaluation
Method
12.78 15.51 15.62 13.88 12.54 14.98 14.69 47
FAHP+TOPSI
S Group
Legal 12.88 15.92 15.47 14.08 12.72 14.82 14.11 118
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
111
differences between the items are small. For this kind of teaching evaluation system,
too large or too small index parameters of each item are not good for the
comprehensive construction of the evaluation system. Therefore, the method used in
this paper with FAHP+TOPSIS combined assessment and genetic algorithm
assignment can better provide suitable index weight values for the evaluation system
of high-quality undergraduate education in private universities. In Table 2, the time
consumption of the four evaluation algorithms for conducting one evaluation is
compared. The time consumed by the method used in this paper, as well as the
combined evaluation of FAHP+TOPSIS and genetic algorithm, is 41s, 38s, 47s, and
118s respectively, and it can be seen that the time consumed by the genetic algorithm
is the largest compared with the other three methods, which is more than twice of the
other methods. For the evaluation system, less elapsed time indicates higher
operational efficiency. Although the method used in this paper, the elapsed time is
about 3s slower compared to the combined FAHP+TOPSIS evaluation. However, the
weights of its seven Level 1 teaching evaluations are more evenly distributed, and the
comprehensiveness of the evaluation system is better than that of the FAHP+TOPSIS
combination evaluation. Because of this, the difference of a few seconds is
acceptable. In summary, the method used in this paper not only has a more even
distribution of index weights but also takes less time to evaluate, which meets our
requirements for the construction of an educational evaluation system.
6. DISCUSSION
For the future of high-quality undergraduate education in private colleges and
universities, it should start from a macro perspective and look into the future. In
response to a greater preference for local regional policy support, it should make
better use of the economic development policies within its region to its advantage.
Within the university, it should be more based on the reality of running schools,
making reliable development plans, and striving to cultivate innovative and high-
quality talents. Not only should they enhance their social responsibility obligations to
increase the brand effect of the school, but they should also build up a high-level,
high-quality faculty from within. It is a continuous process to figure out how to achieve
an effective mechanism and assessment system for high-quality undergraduate
education in private universities. In the establishment of the system, it is a continuous
process of trial and error. In the assessment system, it is a continuous process of
improvement with reasonable methods. The appropriate system and assessment
system are not perfect at the beginning, and the research process should be a
continuous advance.
7. CONCLUSION
In this paper, we analyzed the level 1 teaching evaluation reliability of the samples
based on the deep learning method with 1082 samples collected from the data center
platform of a private university. And subsequently, the method used in this paper was
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
112
compared with three algorithms, namely, FAHP+TOPSIS combined evaluation,
superior order graph method, and genetic algorithm empowerment, and the following
conclusions were obtained:
1. In terms of the distribution of the weight parameters, the difference between
the combined weights of the seven evaluation items is basically small, 12.81%,
15.78%, 15.28%, 14.38%, 12.83%, 12.81%, 15.01%, and 13.27%,
respectively. In terms of individual evaluation items, the difference between the
variable and expected weights of teaching management is more obvious,
reaching 0.90%.
2. In the credibility evaluation, the credibility of the 7 items of the Level 1 teaching
evaluation were 0.89, 0.88, 0.90, 0.91, 0.87, 0.84, and 0.89, respectively. The
calculated results of the sample data of the 7 items were all greater than 0.8,
and all of them could meet the accuracy requirements of the index.
3. Under the comparison of the algorithm of this paper with the combined
assessment of FAHP+TOPSIS, the Euclidean map method, and the genetic
algorithm assigned weights, the seven weight values of the Euclidean map
method differed significantly. Among them, the teaching tools accounted for the
largest weight in the Euclidean map method, reaching 17.34%, and the
teaching style construction accounted for the lowest, 11.78%, with a difference
of 5.56%. The four algorithms took 41 s, 38 s, 47 s, and 118 s. The genetic
algorithm took the most time to assign weights compared to his three methods.
REFERENCES
(1) Neuwirth, L. S., Jović
, S., & Mukherji, B. R. (2021). Reimagining higher
education during and post-COVID-19: Challenges and opportunities. Journal of
Adult and Continuing Education, 27(2), 141-156.
(2) Cong, S. L., Wang, J. J., Li-Yan, X. U., et al. (2018). Discussion on Applied
Transformation of Polymer Professional Undergraduate Education. Polymer
Bulletin.
(3) Murray, H., Walker, M., Dawson, J., et al. (2020). Teaching evidence-based
medicine to medical students using Wikipedia as a platform. Academic Medicine,
95(3), 382-386.
(4) Scott, K. W., Callahan, D. G., Chen, J. J., et al. (2019). Fostering student–faculty
partnerships for continuous curricular improvement in undergraduate medical
education. Academic Medicine, 94(7), 996-1001.
(5) Homberg, A., Klafke, N., Glassen, K., et al. (2020). Role competencies in
interprofessional undergraduate education in complementary and integrative
medicine: a Delphi study. Complementary Therapies in Medicine, 54, 102542.
(6) Marlowe, J., & Tsilomelekis, G. (2020). Accessible and Interactive Learning of
Spectroscopic Parameterization through Computer-Aided Training.
(7) Meng, Y. L., Song, X. Z., Li, Y., et al. (2020). High-Quality Inorganic Chemistry
Teaching During COVID-19. Journal of Chemical Education, 97(9), 2945-2949.
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
113
(8) Xiao, C., Cai, H., Su, Y., et al. (2020). Online teaching practices and strategies
for inorganic chemistry using a combined platform based on DingTalk,
Learning@ ZJU, and WeChat. Journal of Chemical Education, 97(9), 2940-2944.
(9) Zeng, Y. (2020). Evaluation of physical education teaching quality in colleges
based on the hybrid technology of data mining and hidden Markov model.
International Journal of Emerging Technologies in Learning (iJET), 15(1), 4-15.
(10) Cheng, Q., Li, B., & Zhou, Y. (2021). Research on evaluation system of
classroom teaching quality in colleges and universities based on 5G
environment. In Proceedings of the 2021 International Conference on Control
and Intelligent Robotics (pp. 74-85).
(11) Hong, W. (2021). Research on Improving Computer Teaching Quality in Colleges
and Universities in the Context of Big Data. In Journal of Physics: Conference
Series, 1738(1), 012049.
(12) Xu, Z. (2020). Study on the Evaluation of Practical Teaching Quality in Applied
Undergraduate Colleges. In 2020 International Conference on Computing and
Data Science (CDS) (pp. 198-201). IEEE.
(13) Xue, K., Jiang, N., Jiang, H., et al. (2019). Research on the Problems and
Countermeasures of Comprehensive Evaluation of Teachers' Teaching Quality in
Technical Colleges. In 2019 International Conference on Advanced Education,
Management and Humanities (AEMH 2019) (pp. 13-16). Atlantis Press.
(14) Chang, J., Lan, W., & Lan, W. (2021). Higher education innovation and reform
model based on hierarchical probit. Applied Mathematics and Nonlinear
Sciences.
(15) Gao, R., Lloyd, J., & Kim, Y. (2020). A Desirable Combination for Undergraduate
Chemistry Laboratories: Face-to-Face Teaching with Computer-Aided,
Modifiable Program for Grading and Assessment. Journal of Chemical
Education, 97(9), 3028-3032.
(16) Ruder, S. M., & Stanford, C. (2018). Strategies for training undergraduate
teaching assistants to facilitate large active-learning classrooms. Journal of
Chemical Education, 95(12), 2126-2133.
(17) Blythe, A. (2018). Teaching general practice: a rallying flag for undergraduate
education. British Journal of General Practice, 68(677), 560-561.
(18) Ruder, S. M., & Stanford, C. (2020). Training undergraduate teaching assistants
to facilitate and assess process skills in large enrollment courses. Journal of
Chemical Education, 97(10), 3521-3529.
(19) Natarajan, S., Rathinasabapathy, R., Narayanasamy, J., & Aravind, A. R. (2023).
Biometric user authentication system via fingerprints using novel hybrid
optimization tuned deep learning strategy. Traitement du Signal.
(20) Talla-Chumpitaz, R., Castillo-Cara, M., Orozco-Barbosa, L., & Garcia-Castro, R.
(2023). A novel deep learning approach using blurring image techniques for
Bluetooth-based indoor localization. Information Fusion.
(21) Bhor, H. N., & Kalla, M. (2022). Trust-based features for detecting intruders in
the Internet of Things network using deep learning. Computational Intelligence,
38(2), 438-462.
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
114
(22)
Boardman, C., Moschidis, T., & Saunders, M. (2021). EP. TU. 207 Maintaining
High Quality Surgical Teaching during the COVID-19 Pandemic. British Journal
of Surgery, 108(Supplement_7), znab311.029.
(23)
Tackett, S., Zhang, C., Nassery, N., et al. (2019). Describing the evidence base
for accreditation in undergraduate medical education internationally: a scoping
review. Academic Medicine, 94(12), 1995-2008.
(24)
Tempelman, C., Jacobs, U., van der Eijk, M., et al. (2021). Developing a
Catalysis Applied Research Community at a University of Applied Sciences:
From Educational Setting to High-Quality Research Output. Journal of Chemical
Education, 98(8), 2559-2565.
(25)
Perets, E. A., Chabeda, D., Gong, A. Z., et al. (2020). Impact of the emergency
transition to remote teaching on student engagement in a non-STEM
undergraduate chemistry course in the time of COVID-19. Journal of Chemical
Education, 97(9), 2439-2447.
(26)
Wenming, H. (2021). Simulation of English teaching quality evaluation model
based on Gaussian process machine learning. Journal of Intelligent and Fuzzy
Systems, 40(2), 2373-2383.
(27)
Papez, M., & Quinn, A. (2022). Transferring model structure in Bayesian transfer
learning for Gaussian process regression? Knowledge-Based Systems, 251,
108875.
(28)
Hu, C., Ma, Y., & Chen, T. (2021). Application of online process learning
evaluation based on optimal discrete Hopfield neural network and entropy weight
TOPSIS method. Complexity, 2021.
(29)
Kali, Y., Sagy, O., Benichou, M., Atias, O., & Levin-Peled. (2019). Teaching
expertise reconsidered: the technology, pedagogy, content, and space (TPeCS)
knowledge framework. British Journal of Educational Technology, 50(5).
(30)
Nilsson, P. (2020). A framework for investigating qualities of procedural and
conceptual knowledge in mathematics-an inferentialist perspective. Journal for
Research in Mathematics Education, 51(5), 574-599.
https://doi.org/10.17993/3ctic.2023.122.97-115
3C TIC. Cuadernos de desarrollo aplicados a las TIC. ISSN: 2254-6529
Ed.43 | Iss.12 | N.2 April - June 2023
115
