https://doi.org/10.31449/inf.v47i10.5239 Informatica 47 (2023) 53–64 53 
Construction of College English Teaching Effect Evaluation Model 
Based on Artificial Intelligence and Output-Oriented Approach 
Hongjie Li 
Guangzhou Panyu Polytechnic ，Guangzhou City, Guangdong Province 511483 ，China 
E-mail: lihongjie202341@163.com 
Keywords: artificial intelligence, output oriented method, teaching effect evaluation, support vector machine, fuzzy 
comprehensive evaluation  
Received: September 30, 2023 
In order to improve the accuracy and efficiency of teaching effect evaluation results, this paper proposes 
a model design method for college English teaching effect evaluation based on output-oriented method in 
artificial intelligence environment. The standard deviation transformation and range transformation are 
used to standardize the English teaching data, and the fuzzy similarity matrix is established. The data are 
preliminarily classified and clustered according to the results of data standardization; Support vector 
machine is used to classify English teaching data to solve the problem of data imbalance; This paper 
evaluates the effect of college English teaching based on POA theory, follows the construction principle 
of evaluation index, constructs the evaluation index system of college English teaching effect, establishes 
the teaching effect evaluation model based on fuzzy comprehensive evaluation method, and improves the 
traditional fuzzy evaluation model with the idea of minimum membership weighted average deviation 
method to obtain the weight of evaluation index, so as to realize the evaluation of college English teaching 
effect. The experimental results show that the output-oriented method can effectively improve the level of 
English teaching, and the evaluation efficiency of this method is higher and the evaluation results are 
more accurate. 
Povzetek: Predlagan je model za ocenjevanje učinkovitosti poučevanja angleščine z uporabo metode 
umetne inteligence. 
 
1 Introduction 
In recent years, colleges have explored how to 
scientifically evaluate teachers' teaching quality in 
evaluation practice. If products are the life of enterprises, 
teachers' teaching quality is an important guarantee for the 
survival and development of colleges [1, 2]. However, due 
to the fuzziness, diversity and uncertainty of teachers' 
teaching quality evaluation information, as well as the 
influence of various factors such as different perception of 
the importance of various factors in the evaluation 
problem, different personal preferences and different 
knowledge structure by each expert, the actual teachers' 
teaching quality evaluation usually has uncertainty and 
fuzziness. Technically, some evaluation indexes cannot be 
expressed by certain values, it is common to express the 
index value by interval number [3, 4]. However, in the 
actual decision-making process, the decision-making 
results often do not reach a certain degree of consensus 
and are very scattered, resulting in the final evaluation 
results are not convincing. Therefore, it is of certain 
practical significance to study a scientific evaluation 
method of college English teaching effect [5, 6]. 
Reference [7] designed a university theoretical and 
experimental teaching quality evaluation model based on 
RVM machine learning method, improved the kernel 
function and parameter determination method of RVM, 
established a new MK - RVM model for university 
theoretical and experimental teaching quality evaluation  
 
by using the MK - RVM model optimized by genetic 
algorithm, and constructed a result reliability analysis 
method based on confidence interval. The experimental 
results show that the confidence of the evaluation results 
obtained by this method is high, indicating that the 
evaluation results of teaching quality have high reliability, 
but this method has the problem of long evaluation time. 
Reference [8] designed a university teaching quality 
evaluation model based on data mining algorithm. Firstly, 
the model studied and analyzed the relevant literature of 
current university teaching quality evaluation, and 
established the influencing factors of university teaching 
quality evaluation; Then, collect the data of influencing 
factors of college teaching quality, determine the grade of 
college teaching quality through experts, and establish the 
learning sample of college teaching quality evaluation; 
Finally, the BP neural network of data mining technology 
is introduced to train the learning samples to form the 
university teaching quality evaluation model, and the 
advantages of the university teaching quality model are 
analyzed through specific examples. The experimental 
results show that the data mining algorithm can describe 
the differences between the teaching quality grades of 
colleges and universities and obtain high-precision 
teaching quality evaluation results, but the evaluation 
results of this method are subjective and the evaluation 
error of teaching quality is large. Reference [9] designed 
an auxiliary teaching quality evaluation model based on 

54 Informatica 47 (2023) 53–64 H. Li 
active learning support vector machine, comprehensively 
considered various actual situations, and constructed the 
evaluation index system of classroom teaching quality. 
The active learning support vector machine is used to 
establish the classroom teaching quality evaluation model, 
experiment the collected data set related to the teaching 
quality of a university, and analyze the results. The 
experimental results show that the model has certain 
advantages in efficiency and can quickly obtain the 
evaluation results of teaching quality in colleges and 
universities, but the evaluation results are subjective and 
unreliable. Reference [10] proposed a course teaching 
quality evaluation method based on hidden Markov 
algorithm. Taking the learning data of MOOC + SPOC of 
a university learning a professional course as a sample, 
using hidden Markov algorithm and data mining 
technology, the evaluation model of students' learning 
behavior state is established to evaluate students' learning 
behavior in real time. Teachers adjust the teaching content 
according to the evaluation results and interfere with 
students' learning behavior, so as to improve the teaching 
quality. Empirical analysis shows that this method is 
effective. The analysis results have certain reference. 
Through the above analysis, it can be seen that 
although the existing methods have improved the teaching 
evaluation effect to a certain extent and played a certain 
role in promoting the improvement of teaching quality, 
due to the large number of teaching data, the diversity of 
teaching objects and other reasons, there will be some 
errors in the teaching effect evaluation and do not have 
high efficiency. In order to improve the accuracy and 
reliability of college English teaching effect evaluation, 
with the support of artificial intelligence technology, this 
paper puts forward a college English teaching effect 
evaluation model based on output-oriented method in 
artificial intelligence environment, in order to provide a 
new and effective way for college teaching quality 
evaluation. 
2 Application of artificial intelligence 
technology in the evaluation of 
college English teaching effect 
With the development of machine learning [11, 12] and 
big data, how to use artificial intelligence technology to 
optimize teaching process and improve teaching 
evaluation has become a common topic of concern for 
education authorities, researchers, education technology 
companies and educators. In recent years, billions of 
learners carry out formal and informal learning anytime 
and anywhere on various learning platforms, forming a 
specific activity track and a large amount of learning data. 
The application of artificial intelligence technology to 
analyze the massive learning data in the digital learning 
environment and provide automatic feedback and 
evaluation to students has been widely recognized. 
Therefore, using artificial intelligence technology [13] and 
big data analysis to improve the efficiency and 
effectiveness of education evaluation has also attracted 
more and more attention of researchers. In the above 
context. In this paper, data mining, support vector machine 
and other technologies are applied to the evaluation of 
English teaching effect, and the teaching data are 
processed scientifically, so as to improve the efficiency 
and accuracy of teaching effect evaluation. 
2.1 English teaching data mining 
After years of school running, a school must have a large 
amount of classroom teaching evaluation data. These data 
have become a valuable resource for school management 
and decision-making. Finding valuable information from 
these data has become a very arduous task. In fact, it is 
impossible and unnecessary to classify these data 
accurately in the process of mining. Practical 
classification is often accompanied by fuzziness, so using 
fuzzy theory to cluster analysis, and then teaching 
evaluation, will appear more natural and more in line with 
the objective facts [14]. The specific steps of data mining 
using fuzzy theory are as follows: 
2.1.1 Data standardization 
In the case of collecting a large amount of classroom 
teaching quality data, establish the overall data records of 
the teaching evaluation data warehouse, and establish the 
sample sets 𝑆 to be classified, 𝑆 = { 𝑠 1
, 𝑠 2
, . . . , 𝑠 𝑛 }, where 
𝑛 represents the number of samples. The quantified 
attribute data of specific qualitative teaching evaluation 
indicators are called sample indicators, and there are 𝑀 
indicators. This can be described by an 𝑚 -dimensional 
vector, which is represented by a set as 𝑆 𝑖 =
{ 𝑠 𝑖 1
, 𝑠 𝑖 2
, . . . , 𝑠 𝑖𝑚
}, and the original data matrix is obtained 
as: 
𝑆 = [
𝑠 11
𝑠 12
. . . 𝑠 1 𝑛 𝑠 21
𝑠 22
. . . 𝑠 2 𝑛 ⋮ ⋮ . . . ⋮
𝑠 𝑛 1
𝑠 𝑛 2
. . . 𝑠 𝑛𝑚
] (1) 
Since the collected data is often not the data of [0,1] 
closed interval, according to the requirements of fuzzy 
matrix, the data transformation is compressed to the 
interval [0,1] through standard deviation transformation 
and range change. 
(1) Standard deviation transformation 
𝑠 𝑖𝑙
′
= √ 𝑠 𝑖𝑙
− 𝑠 𝑖𝑙
 
(2) 
Among them: 𝑖 = 1 , 2 , . . . , 𝑛 , 𝑙 = 1 , 2 , . . . , 𝑚 , 𝑠 𝑖𝑙
 
represents the standard deviation, 𝑠 𝑖𝑙
 represents the mean 
value. 
The normalized data 𝑠 𝑖𝑙
′
 obtained at this time is not 
necessarily within the [0,1] interval, therefore, the 
following transformation is required. 
(2) range transformation 

Construction of College English Teaching Effect Evaluation Model… Informatica 47 (2023) 53–64 55 
( ) ( )
''
max ' min '
11
il il
il
i n il i n il
ss
s
ss
   
−
=
−
 
(3) 
2.1.2 Establish a fuzzy similarity matrix 
In teaching data mining, in order to improve the efficiency 
of data mining, the data will be preliminarily classified 
according to the results of data standardization [15]. In this 
process, a statistic ℎ
𝑖𝑗
 is obtained by measuring the 
similarity between the classified objects. Set up a universe 
of discourse 𝑊 = { 𝑤 1
, 𝑤 2
, . . . , 𝑤 𝑛 }, where each element is 
a sample, and establish a fuzzy similarity matrix on 𝑊 : 
𝐻 𝑖𝑗
= [
ℎ
11
ℎ
12
. . . ℎ
1 𝑛 ℎ
21
ℎ
22
. . . ℎ
2 𝑛 ⋮ ⋮ . . . ⋮
ℎ
𝑛 1
ℎ
𝑛 2
. . . ℎ
𝑛𝑚
] (4) 
Among them: 0 ≤ ℎ
𝑖𝑗
≤ 1, 𝑖 = 1 , 2 , . . . , 𝑛 , 𝑗 =
1 , 2 , . . . , 𝑛 . 
There are many ways to calculate ℎ
𝑖𝑗
, this paper uses 
the number product method to calculate it: 
ℎ
𝑖𝑗
= ∑ 𝑠 𝑖𝑘
+ 𝑠 𝑗𝑘
𝑚 𝑘 = 1
 (5) 
Among them: 𝑠 𝑖𝑘
 and 𝑠 𝑗𝑘
 both represent the vector 
dot product in the high-dimensional feature space, and 𝑘 
represents the selected constant, which satisfies the 
following conditions: 
min
1
m
i j ik jk
k
k s s

=
+

 
(6) 
2.1.3 Cluster analysis 
According to formula (4), the standardized data is 
clustered and analyzed. The commonly used clustering 
analysis methods mainly include the transitive closure 
method, the maximum tree method and the netting 
method. The transitive closure method is relatively easy to 
implement with a computer. This method generates a 
clustering graph based on the similarity matrix. The 
fuzziness of the clustering process is mainly reflected in 
the similarity matrix, which can be measured by fuzzy 
information entropy, which is beneficial to improve the 
efficiency of data mining. Therefore, use the square 
method to find the transitive closure 𝐻 𝑖𝑗
( 𝑥 ) of the fuzzy 
similarity matrix 𝐻 𝑖𝑗
: 
𝐻 𝑖𝑗
( 𝑥 ) = 𝐻 𝑖𝑗
𝑘 − 𝑅 𝑖𝑗
𝑘 (7) 
Among them: 𝐻 𝑖𝑗
𝑘 and 𝑅 𝑖𝑗
𝑘 both represent fuzzy 
information entropy. This forms a dynamic clustering 
diagram of the teaching data, as shown in Figure 1. 
 
 
1 2 3 4 5 6 0
1
2
3
4
5
6
X axis
Y axis
 
Figure1: Data dynamic clustering diagram 

56 Informatica 47 (2023) 53–64 H. Li 
 
2.2 Classification of English teaching data 
Because the teaching process is complex and there are 
many types of data, data classification processing is 
carried out on the basis of data mining. Support vector 
machine has the functions of classification, prediction, 
clustering, anomaly detection and density estimation. It is 
a general learning machine. As one of the mainstream 
technologies of machine learning, the English teaching 
data classification method constructed according to this 
technical scheme has good universality, so this method is 
selected for data classification processing [16]. 
2.2.1 Support vector machine 
Support vector machine belongs to the category of 
computational learning in artificial intelligence. It is a 
machine learning technology developed in the mid-1990s. 
Compared with traditional learning technology, it has a 
solid theoretical foundation. From a large number of 
applications, its performance is superior. It has created and 
maintained the best records on many specific problems 
such as handwritten numeral recognition and text 
classification, involving pattern recognition, regression 
estimation, data mining, information retrieval Intelligent 
signal processing, intelligent control, nonlinear mode 
reconstruction and other fields [17, 18]. 
The theoretical basis of support vector machine is 
statistical learning theory, which has a strict mathematical 
basis [19]. The research of this theory began in the late 
1960s. In the past 30 years, although a large number of 
scholars have done pioneering and foundational work, the 
work in this period is mainly purely theoretical. According 
to this principle, it cannot fit the known training data well, 
and the generalization ability is not strong. Therefore, 
based on the existing support vector machine theory, 
support vector machine with quadratic penalty term is 
used to classify English teaching data [20, 21]. 
Let 𝑦 𝑘 represent the input vector, 𝐸 is the feature 
space, in which, there are a large number of feature vectors 
𝑒 𝑘 ∈ 𝐸 , 𝜙 is the feature map, then there is a kernel 
function 𝐾 ( 𝑦 𝑘 ): 
𝐾 ( 𝑦 𝑘 ) = 𝜙 ( 𝑦 𝑘 ) × 𝐺 𝑥 (8) 
Among them: 𝐺 𝑥 represents the inner product in 𝐸 . 
Let { 𝑓 𝑎𝑏
}
𝑘 be the training set, where 𝑎 and 𝑏 are the set of 
indicators for class 1 and 2, respectively, 𝑎 ∪ 𝑏 =
{ 1 , 2 , . . . , 𝑚 }. The optimization problem associated with 
support vector machines when 𝑒 𝑘 is linearly separable is: 
𝑚 𝑖 𝑛 1
2
‖ 𝑒 ‖
2
 (9) 
Among them: 
{
𝑤 ⋅ 𝑒 𝑖 + 𝑑 ≥ 1
∀ 𝑎 ∈ 𝐴 𝑤 ⋅ 𝑒 𝑗 + 𝑑 ≤ − 1
 (10) 
When 𝑒 𝑘 is nonlinearly separable, a quadratic 
penalty is used to optimize the support vector machine: 
𝑚 𝑖 𝑛 1
2
‖ 𝑒 ‖
2
− ∑ 𝜂 𝑘 2
𝑚 𝑘 = 1
 (11) 
Among them: 
{
𝑤 ⋅ 𝑒 𝑖 + 𝑑 ≥ 1 − 𝜂 𝑘 2
∀ 𝑎 ∈ 𝐴 𝑤 ⋅ 𝑒 𝑗 + 𝑑 ≤ − 1 + 𝜂 𝑘 2
 (12) 
Let 𝜆 𝑖 be the solution of the above dual problem, if 
𝜆 𝑖 ≠ 0, call the corresponding 𝑦 𝑖 a support vector, then the 
discriminant function is called a support vector machine. 
The support vector machine is characterized by 
training samples and kernel functions. The schematic 
diagram of the support vector machine is shown in Figure 
2. 
( )
k
Ky
( )
1
Ky
1
a
2
a
n
a
 
Figure 2: Schematic diagram of the structure of support vector machine 

Construction of College English Teaching Effect Evaluation Model… Informatica 47 (2023) 53–64 57 
2.2.2 Classification of English teaching data 
based on support vector machine 
According to the above theory, given a sample set ( 𝑥 𝑖 , 𝑦 𝑖 ), 
a hyperplane 𝑤 ⋅ 𝜙 ( 𝑥 ) + 𝑓 = 0 is used to divide two types 
of samples, and its classification function can be expressed 
as: 
𝑓 ( 𝑥 ) = 𝑠 𝑔𝑛 [ 𝑤 ⋅ 𝜙 ( 𝑥 ) + 𝑓 ] (13) 
Among them: 𝜙 ( 𝑥 ) maps the sample set to a high-
dimensional space to make it linearly separable. The goal 
of SVM is to determine 𝑤 and 𝑓 according to the training 
samples, so as to maximize the classification interval 
between the two types of samples [22, 23]. Standard SVM 
determines 𝑤 and 𝑓 by solving the optimization problem 
in Equation (13): 
𝑚 𝑖 𝑛 𝑍 ( 𝑤 , 𝑓 ) = 𝑍 𝜂 𝑖 +
1
2
‖ 𝑓 ‖
2
+ 𝐷 1
2
( ∑ 𝜙 𝑖 𝑛 𝑖 = 1
) (14) 
Among them: In the objective function, ‖ 𝑓 ‖
2
 
represents the reciprocal of the interval between the two 
types of samples, 𝐷 represents the penalty factor, which 
controls the balance between the empirical error and the 
classification interval, and 𝜂 𝑖 represents the training error 
of the 𝑖 -th sample. Since equation (14) is a convex 
quadratic optimization problem [24, 25] it can be 
transformed into its dual problem to solve: 
𝑚 𝑎𝑥 𝑊 ( 𝜀 ) = ∑ 𝜀 𝑖 𝑛 𝑖 = 1
−
1
2
∑ 𝜀 𝑖 𝜀 𝑗 𝛽 𝑖 𝛽 𝑗 𝑛 𝑖 = 1
 (15) 
Among them: 𝜀 𝑖 and 𝜀 𝑗 represent the finite kernel 
function; 𝛽 𝑖 and 𝛽 𝑗 represent the quadratic objective 
function. 
Solving Equation (15) yields the optimal 𝜀 𝑖 , 
therefore, 𝑤 is expressed in the form of Equation (16): 
𝑤 = ∫ 𝜀 𝑖 𝑛 𝑖 = 1
𝛽 𝑖 𝜙 𝑖 (16) 
Convert the classification function to: 
𝑓 ( 𝑥 ) = 𝑠 𝑔𝑛 ∑ ∑ 𝜀 𝑖 𝛽 𝑖 𝜙 𝑖 𝑚 𝑖 = 1
𝑛 𝑖 = 1
 (17) 
Since it is very difficult to perform dot product 
operations in high-dimensional space, sometimes it is 
even impossible, so SVM uses the idea of the kernel 
function, and uses the kernel function 𝐾 ( 𝑥 𝑖 , 𝑦 𝑖 ) = 𝜙 ( 𝑥 𝑖 ) ⋅
𝜙 ( 𝑥 𝑗 ) that satisfies the Mercer condition to replace the 
high-dimensional space dot product operation. Therefore, 
the classification function is transformed into is formula 
(18): 
𝑓 ( 𝑥 ) = 𝑠 𝑔𝑛 ∑ ∑ 𝐾 ( 𝑥 𝑖 , 𝑦 𝑖 ) + 𝑓 𝑚 𝑖 = 1
𝑛 𝑖 = 1
 (18) 
SVM can achieve better classification effect under 
the condition of balanced distribution of sample set. 
However, when the sample set is unbalanced, the 
classification effect of a small number of samples is poor. 
At present, the processing methods for unbalanced data 
classification can be roughly divided into three categories. 
One is oversampling a small number of samples, that is, 
increasing the number of small samples through 
replication or interpolation, but this method follows a 
hypothesis: There are still many classes between multi 
class samples and few classes near few class samples. This 
assumption is related to data characteristics and 
sometimes does not hold. Therefore, this method is highly 
dependent on data characteristics. The second category is 
the use of multiple sample pruning methods, that is, the 
number of samples is reduced. The third kind of method is 
the algorithm level method, which mainly includes 
moving the classification hyperplane in the high-
dimensional feature space, assigning different penalty 
coefficients to different categories of samples, and 
modifying the form of kernel function according to the 
characteristics of samples. According to the characteristics 
of many types and large quantities of college English 
teaching data, this paper selects the third method to further 
process the classified data in order to solve the problem of 
data imbalance [26]. 
3 Model design of college English 
teaching effect evaluation based on 
output-oriented method 
3.1 Output-oriented approach 
Output oriented approach (POA) is an advanced teaching 
method with strong practical application function. It is a 
bold attempt to the classroom teaching mode and has 
achieved good teaching results. POA theory, namely 
output oriented method, mainly includes three main links: 
teaching theory, teaching hypothesis and teaching process, 
which are connected and supported by each other. POA 
theory emphasizes teaching output, that is, it is necessary 
to clarify the teaching results you want to obtain before the 
beginning of teaching activities, organize teaching ideas 
and assumptions centered on learning, design teaching 
processes and promote teaching development. 
The theoretical system of Poa consists of three parts, 
among which "teaching concept" is the guiding ideology 
of the other two parts; "Teaching hypothesis" is the 
theoretical support of "teaching process"; "Teaching 
process" is the realization of "teaching concept" and 
"teaching hypothesis". At the same time, the intermediary 
role of teachers is reflected in all links of "teaching 
process". 

58 Informatica 47 (2023) 53–64 H. Li 
3.2 Design of evaluation model of college 
English teaching effect 
This paper evaluates the effect of college English teaching 
based on POA theory, and makes an objective evaluation 
of the effectiveness of POA. In this paper, the fuzzy 
comprehensive evaluation method is used to design the 
evaluation model of college English teaching effect. 
Unlike the traditional comprehensive evaluation method, 
this method can be expressed by some simple numerical 
values, and then use the sum or weighted average method 
to get a total score and sort the best. Instead, a complete 
evaluation index system is constructed, and the evaluation 
index weight is solved for this system. According to the 
weight calculation results, the evaluation results of college 
English teaching effect are obtained. According to the 
characteristics of college English classroom teaching, this 
paper designs the following evaluation process of college 
English teaching effect. 
3.2.1 Principles of teaching effect evaluation 
The establishment of a college English teaching effect 
evaluation system should follow the following principles: 
(1) The principle of scientific orientation: The 
ultimate goal of conducting classroom teaching quality 
assessment is to improve teaching quality. 
(2) Consistency principle: The evaluation index 
should be adapted to the current teaching management 
system of colleges. 
(3) Integrity principle: The evaluation indicators 
should comprehensively and comprehensively reflect the 
teaching objectives, and provide comprehensive teaching 
information for teaching management. 
(4) Independence principle: each index is 
independent of each other, neither will it affect the 
scientific of the evaluation results due to overlapping 
indicators, nor will it increase the evaluation workload and 
reduce the feasibility of evaluation due to redundant 
indicators. 
(5) The principle of plasticity: teachers can clearly 
recognize their own strengths and weaknesses through 
evaluation indicators and evaluation results, so as to take 
targeted measures to improve their work. 
(6) The principle of operability: The design of 
indicators adopts a combination of quantitative and 
qualitative methods. When expressing the teaching effect 
evaluation indicators, try to be as clear and unambiguous 
as possible. At the same time, try to be detailed in the 
design. The connotation of the indicators requires that 
there be realistic and available materials as a basis, and can 
be measured objectively, so as to minimize the errors 
caused by subjectivity. 
3.2.2 Evaluation index system 
Based on the above principles, the evaluation index 
system of college English teaching effect is constructed, 
as shown in Table 1
 
Table 1: The evaluation index system of college English teaching effect 
The evaluation 
index of college 
English 
teaching effect 
First-level indicator Secondary indicators 
S 1 teaching attitude 
S 11 Be serious and responsible, be a teacher 
S 12 Seriously teach, not perfunctory 
S 13 Rigorous teaching and self-discipline 
S 2 teaching method 
S 21 Flexibility and variety of teaching methods 
S 22 Reasonable allocation of class hours and moderate progress 
S 23 Emphasis and difficulty 
S 24 Focus on cultivating students' language communicative ability 
S 3 language of instruction 
S 31 pronunciation standard 
S 32 well organized 
S 33 language specification 
S 4 Teaching effect 
S 41 Active classroom atmosphere 
S 42 Students are highly cooperative 
S 43 Classroom is in good order 
S 44 After-school review effect 
S 45 Excellent test scores 
3.2.3 Evaluation model design 
Fuzzy comprehensive evaluation method is an evaluation 
method that uses fuzzy mapping and fuzzy linear 
transformation to make comprehensive decisions. 
Because college English teaching is a complex process of 
information exchange between teachers and students, it 
has the characteristics of multiple factors, ambiguity, 
effectiveness and dynamics, using the fuzzy 
comprehensive evaluation method as the basis to establish  
 
a teaching effect evaluation model, and using the idea of 
the minimum membership degree weighted average 
deviation method to improve the traditional fuzzy 
evaluation model, the improved multi-layer fuzzy 
comprehensive evaluation steps are as follows: 
(1) Determine the factor set 𝑍 = { 𝑧 1
, 𝑧 2
, . . . , 𝑧 𝑛 } of 
the evaluation object, where 𝑧 𝑖 = { 𝑧 𝑖 1
, 𝑧 𝑖 2
, . . . , 𝑧 𝑖𝑚
} is the 𝑖 
-th factor in the upper layer determined by the 𝑚 factors 
in the lower layer. 

Construction of College English Teaching Effect Evaluation Model… Informatica 47 (2023) 53–64 59 
(2) Establish an evaluation set 𝐽 = { 𝑗 1
, 𝑗 2
, . . . , 𝑗 𝑚 }. 
(3) Establish the single-factor evaluation matrix 𝐸 𝑖 
of the next layer: 
𝐸 𝑖 = [
𝑒 𝑖 11
𝑒 𝑖 12
. . . 𝑒 𝑖 1 𝑙 𝑒 𝑖 21
𝑒 𝑖 22
. . . 𝑒 𝑖 2 𝑙 ⋮ ⋮ ⋱ ⋮
𝑒 𝑖𝑛 1
𝑒 𝑖𝑛 2
. . . 𝑒 𝑖 𝑛 𝑝 ] (19) 
The weight set is obtained according to formula (19): 
𝑉 𝑖 = { 𝑣 𝑖 1
, 𝑣 𝑖 2
, . . . , 𝑣 𝑖𝑛
} (20) 
Among them: ∑ 𝑣 𝑖 𝑛 𝑖 = 1
= 1. 
(4) Carry out a comprehensive evaluation of the 
lower level 
𝐶 𝑖 = 𝑉 𝑖 × 𝐸 𝑖 (21) 
(5) Carry out the fuzzy comprehensive evaluation of 
the upper layer, and form the upper layer fuzzy matrix 
according to the comprehensive evaluation results of the 
sub-targets obtained in step (4): 
𝐸 = [
𝑐 1
𝑐 2
⋮
𝑐 𝑚 ] (22) 
According to formula (22), the weight corresponding 
to the index of this layer is obtained, and then the 
comprehensive evaluation result of this layer is obtained. 
(6) Calculate the comprehensive evaluation value, 
and use the double weight method to convert the above 
comprehensive evaluation results into the corresponding 
comprehensive evaluation value, thereby obtaining the 
specific weight of each evaluation index. 
3.2.4 Weights of evaluation indicators 
According to the above steps, the specific weights of the 
evaluation indicators of college English teaching effect are 
calculated, and the results are shown in Table 2. 
 
Table 2: Weights of evaluation indicators of college English teaching effect 
The evaluation index of 
college English teaching 
effect 
First-level indicator Secondary indicators 
S 1 teaching attitude 0.23 
S 11 Be serious and responsible, be a teacher 0.37 
S 12 Seriously teach, not perfunctory 0.33 
S 13 Rigorous teaching and self-discipline 0.30 
S 2 teaching method 0.31 
S 21 Flexibility and variety of teaching methods 0.15 
S 22 Reasonable allocation of class hours and moderate 
progress 0.22 
S 23 Emphasis and difficulty 0.34 
S 24 Focus on cultivating students' language 
communicative ability 0.29 
S 3 language of 
instruction 0.17 
S 31 pronunciation standard 0.34 
S 32 well organized 0.29 
S 33 language specification 0.37 
S 4 Teaching effect 0.29 
S 41 Active classroom atmosphere 0.12 
S 42 Students are highly cooperative 0.19 
S 43 Classroom is in good order 0.20 
S 44 After-school review effect 0.21 
S 45 Excellent test scores 0.28 
 
 
4 Evaluation effect research 
In order to verify the effectiveness of the evaluation model 
of college English teaching effect based on the output-
oriented method in the artificial intelligence environment, 
the application effect is tested by a combination of 
qualitative and quantitative methods, and the teaching 
quality is evaluated and analyzed. 
4.1 Scheme design 
(1) Research objects 
This study took students from two teaching classes 
(Class A and Class B) of a university in 2018 as the 
research objects, with 36 students in each class. 
 
(2) research proposal 
This study takes one semester as the investigation 
period, and the investigation object is mainly the teaching 
class of students. English teaching is a whole and organic 
system. Therefore, we should regard the distribution of 
each English achievement of the whole teaching class as a 
developing complete system, and regard each student's 
achievement as the result of a random experiment. The 
specific test methods are as follows: in the process of 
quantifying the effect index of English teaching, based on 
two tests, divide the score of an investigation object from 
high to low into eight grades, and then calculate the 
proportion of the number of students at each grade in the 
total number; After a certain stage, continue to analyze the 

60 Informatica 47 (2023) 53–64 H. Li 
grade changes of each of the above students, so as to 
obtain the effect of English teaching. 
(3) Research methods 
In this study, SPSS 15.0 statistical analysis software 
was used for data processing of the English scores of 
students in two classes. 
4.2 Analysis of results 
Table 3 and Table 4 are the changes in the scores of 
students in classes A and B before and after the output-
oriented English teaching. The traditional method is to 
divide the grades of students in a class into five grades: 
excellent, good, medium, pass and fail. Since students' 
physical education scores are normally distributed, 
students' scores should be divided into eight grades, 
namely, excellent, excellent, good, medium, pass, fail, 
poor, and very poor.
 
Table 3: Before teaching the output-oriented method 
class excellent very good good medium pass fail poor very poor 
A 2 5 4 15 5 3 2 0 
B 1 7 5 10 8 2 3 0 
 
Table 4: After teaching the output-oriented method 
class excellent very good good medium pass fail poor very poor 
A 5 9 7 6 6 2 1 0 
B 3 10 8 5 5 4 1 0 
 
 
It can be seen from the data in Table 3 and Table 4 
that after using the output-oriented method, students' 
English scores have been significantly improved, which is 
specifically reflected in the increase in the number of high-
level students, indicating that the output-oriented method 
can effectively improve the quality of college English 
teaching. This is because the output-oriented method takes 
into account the students' original English foundation and 
pays attention to the investigation of the students' progress 
in English learning. It can effectively mobilize the 
students' enthusiasm in English class and give full play to 
the students' main role in English learning. According to 
the above experimental results, it can be concluded that 
this method can scientifically and objectively evaluate the 
effect of college English teaching, which is conducive to 
teachers' teaching feedback and teaching summary, 
continuously improve the teaching level, give full play to 
the leading role of teachers in the teaching process, and 
better improve the teaching quality. 
(2) Application effect analysis 
In order to further verify the effectiveness of this 
method, reference [7] method and reference [8] method 
are used as comparison methods to compare the evaluation 
effects of different methods from the perspective of 
evaluation efficiency and evaluation error. The 
comparison results of evaluation efficiency are shown in 
Figure 3 and Figure 4.
0
Evaluation time/min
2
4
6
8
10
The method of this paper
Reference [7] method
2 3 4 5 6
Number of experiments/time
1
Reference  [8] method
 
Figure 3: Comparison results of evaluation efficiency of class A 
 
 

Construction of College English Teaching Effect Evaluation Model… Informatica 47 (2023) 53–64 61 
0
Evaluation time/min
2
4
6
8
10
The method of this paper
Reference  [7] method
2 3 4 5 6
Number of experiments/time
1
Reference  [8] method
 
Figure 4: Comparison results of evaluation efficiency of class B 
 
 
From the data in Figure 3 and Figure 4, it can be seen 
that during the experiment, the evaluation time required 
for the evaluation of teaching effect by using the method 
of reference [7] and reference [8] is higher than 4 minutes, 
while in the same experimental environment, the 
evaluation time required for the teaching effect evaluation 
using the method in this paper is all less than 2.5min. It 
can be seen that the method in this paper effectively 
improves the efficiency of teaching effect evaluation, and 
also improves its application performance in practical 
work. 
Figure 5 and Figure 6 show the comparison results 
of teaching effect evaluation errors of different methods in 
class A and class B respectively.
 
.
0
Evaluation error/%
2
4
6
8
10
The method of this paper
Reference  [7] method
2 3 4 5 6
Number of experiments/time
1
Reference  [8] method
 
Figure 5: Comparison results of evaluation errors of class A 
 
 
 

62 Informatica 47 (2023) 53–64 H. Li 
0
Evaluation error/%
2
4
6
8
10
The method of this paper
Reference [7] method
2 3 4 5 6
Number of experiments/time
1
Reference [8] method
 
Figure 6: Comparison results of evaluation errors of class B 
 
It can be seen from Figure 5 and Figure 6 that the 
teaching effect evaluation error gap between the method 
of reference [7] and the method of reference [8] is small 
and basically maintained at the same level, while the 
teaching effect evaluation error of this method is 
significantly lower than that of the method of reference [7] 
and the method of reference [8]. It shows that the 
evaluation result of this method is more reliable and 
feasible. 
5 Conclusion 
Under the environment of artificial intelligence, this paper 
puts forward the design method of college English 
teaching effect evaluation model based on output-oriented 
method, establishes a set of index system suitable for 
college English teaching effect evaluation, and introduces 
the methods of data standardization and data classification 
into college English teaching effect evaluation. The 
experimental results show that this method has the 
advantages of high evaluation efficiency and more 
accurate evaluation results. This is because this method 
fully considers the fuzziness of English teaching 
information and makes the evaluation results closer to the 
objective reality. This study not only provides the 
corresponding basis for the guidelines and decision-
making of the education management department, but also 
provides a direction worthy of thinking and efforts for 
teachers to improve their teaching level. 
 
Data Availability 
The raw data supporting the conclusions of this article wi
ll be made available by the authors, without undue reserv
ation." 
Conflicts of Interest 
The authors declared that they have no conflicts of intere
st regarding this work." 
 
 
Funding 
The work is not supported by any funding. 
Authorship contribution statement 
Hongjie Li: Writing-Original draft preparation, 
Conceptualization, Supervision, Project administration. 
Declarations 
Not applicable 
Refrences 
 [1]  Noben I, Deinum JF, Hofman WHA (2022) Quality 
of teaching in higher education: reviewing teaching 
behaviour through classroom observations. 
International Journal for Academic Development 
27:31–44. 
https://doi.org/10.1080/1360144X.2020.1830776 
 [2]  Veza O (2021) Quality of teacher learning 
implementation plan and its implementation in 
preparing vocational school students for ready to 
work. Technical and Vocational Education 
International Journal (TAVEIJ) 1:41–48 
 [3]  Putri NR, Sari FM (2021) Investigating English 
teaching strategies to reduce online teaching 
obstacles in the secondary school. Journal of 
English Language Teaching and Learning 2:23–31 
 [4]  Sun Z, Anbarasan M, Praveen Kumar D (2021) 
Design of online intelligent English teaching 
platform based on artificial intelligence techniques. 
Comput Intell 37:1166–1180. 
https://doi.org/10.1111/coin.12351 
 [5]  Rozal E, Ananda R, Zb A, et al (2021) The effect of 
project-based learning through YouTube 
presentations on English learning outcomes in 
physics. AL-Ishlah: Jurnal Pendidikan 13:1924–
1933. Doi:10.35445/alishlah. v13i3.1241 
 [6]  Zou C, Li P, Jin L (2021) Online college English 
education in Wuhan against the COVID-19 

Construction of College English Teaching Effect Evaluation Model… Informatica 47 (2023) 53–64 63 
pandemic: Student and teacher readiness, 
challenges and implications. PLoS One 16: 
e0258137. 
https://doi.org/10.1371/journal.pone.0258137 
 [7]  Wei C, Tsai S-B (2022) Evaluation model of college 
English teaching effect based on particle swarm 
algorithm and support vector machine. Math Probl 
Eng 2022: https://doi.org/10.1155/2022/7132900 
 [8]  Li Y (2020) Design of university teaching quality 
evaluation model based on data mining algorithm. 
Modern Electronics Technique 43:119–122 
 [9]  Zhang Y (2019) Assisted teaching quality 
evaluation model based on active learning support 
vector machine. Modern Electronics Technique 
42:112–114 
 [10]  Xiao LP, Tian ZJ, Tian JW (2022) Research on 
curriculum teaching quality evaluation based on 
hidden Markov algorithm. Heilongjiang Higher 
Education Research 40:151–155 
 [11]  D’Amour A, Heller K, Moldovan D, et al (2022) 
Underspecification presents challenges for 
credibility in modern machine learning. The Journal 
of Machine Learning Research 23:10237–10297. 
Doi:10.5555/3586589.3586815 
 [12]  Greener JG, Kandathil SM, Moffat L, Jones DT 
(2022) A guide to machine learning for biologists. 
Nat Rev Mol Cell Biol 23:40–55. 
https://doi.org/10.1038/s41580-021-00407-0 
 [13]  Lee D, Yoon SN (2021) Application of artificial 
intelligence-based technologies in the healthcare 
industry: Opportunities and challenges. Int J 
Environ Res Public Health 18:271. 
https://doi.org/10.3390/ijerph18010271 
 [14]  Vrontis D, Christofi M, Pereira V, et al (2022) 
Artificial intelligence, robotics, advanced 
technologies and human resource management: a 
systematic review. The International Journal of 
Human Resource Management 33:1237–1266. 
https://doi.org/10.1080/09585192.2020.1871398 
 [15]  Dogan A, Birant D (2021) Machine learning and 
data mining in manufacturing. Expert Syst Appl 
166:114060. 
https://doi.org/10.1016/j.eswa.2020.114060 
 [16]  Laoli A, Dakhi O, Zagoto MM (2022) The 
Application of Lesson Study in Improving the 
Quality of English Teaching. Edukatif: Jurnal Ilmu 
Pendidikan 4:2238–2246. 
https://doi.org/10.31004/edukatif.v4i2.2434 
 [17]  Li N (2021) A fuzzy evaluation model of college 
English teaching quality based on analytic hierarchy 
process. International Journal of Emerging 
Technologies in Learning (iJET) 16:17–30. 
https://doi.org/10.3991/ijet.v16i02.19731 
 [18]  Syafiq AN, Rahmawati A, Anwari A, Oktaviana T 
(2021) Increasing speaking skill through YouTube 
video as English learning material during online 
learning in pandemic covid-19. Elsya: Journal of 
English Language Studies 3:50–55. 
https://doi.org/10.31849/elsya.v3i1.6206 
 [19]  Li M, Zhang Q, Yu Z (2021) Integrating educational 
games into college English learning. Educ Res Int 
2021:1–9. https://doi.org/10.1155/2021/6403034 
 [20]  Kurani A, Doshi P, Vakharia A, Shah M (2023) A 
comprehensive comparative study of artificial 
neural network (ANN) and support vector machines 
(SVM) on stock forecasting. Annals of Data Science 
10:183–208. https://doi.org/10.1007/s40745-021-
00344-x 
 [21]  Zhang D, Perez-Paredes P (2021) Chinese 
postgraduate EFL learners’ self-directed use of 
mobile English learning resources. Comput Assist 
Lang Learn 34:1128–1153. 
https://doi.org/10.1080/09588221.2019.1662455 
 [22]  Álvarez-Alvarado JM, Ríos-Moreno JG, Obregón-
Biosca SA, et al (2021) Hybrid techniques to predict 
solar radiation using support vector machine and 
search optimization algorithms: a review. Applied 
Sciences 11:1044. 
https://doi.org/10.3390/app11031044 
 [23]  Nurkholis A, Abidin Z, Sulistiani H (2021) 
Optimasi Parameter Support Vector Machine 
Berbasis Algoritma Firefly Pada Data Opini Film. 
Jurnal RESTI (Rekayasa Sistem Dan Teknologi 
Informasi) 5:904–910. 
https://doi.org/10.29207/resti.v5i5.3380 
 [24]  Liang K, Hao Y, Zhao J (2021) An empirical study 
on the effect and quality of situational English 
teaching. International Journal of Emerging 
Technologies in Learning (iJET) 16:74–86. 
https://doi.org/10.29207/resti.v5i5.3380 
 [25]  Putra RWP (2021) Improving the students’ 
motivation in learning English through Google 
Meet during the online learning. English Learning 
Innovation (englie) 2:35–42. 
https://doi.org/10.22219/englie.v2i1.14605 
 [26]  Marasri S, Thongdee V, Homsombat P, et al (2021) 
CLM English Teaching and Learning Guidelines 
for Undergraduate Students in Social Studies at 
Phutthachinnarat Sangha College, MCU. Online 
Submission 5:277–286. 
https://doi.org/10.21744/lingcure.v5nS3.1606 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
 
 

64 Informatica 47 (2023) 53–64 H. Li