Construction and Practical Exploration of Intelligent Teaching Evaluation System in Higher Vocational Colleges and Universities

With the rapid development of science and technology, the application of intelligent technology in the field of education is becoming more and more widespread. Intelligent teaching evaluation system is a system that utilizes artificial intelligence, big data, cloud computing and other technical means to comprehensively assess and analyze the teaching process and results in colleges and universities. Intelligent teaching evaluation system in colleges and universities, as a tool for comprehensive and scientific assessment of teaching quality, is of great significance in college and university education [1-4].

Intelligent education evaluation system has the following characteristics: First, diversified evaluation: intelligent education evaluation system takes into account students’ academic performance, actual performance, participation and other aspects of the data, in order to comprehensively understand the degree of mastery and learning. Second, real-time feedback: intelligent education evaluation system can monitor students’ learning progress in time, provide real-time feedback on students’ learning, and help students adjust their learning strategies in time [5-8]. Third, personalized assessment: intelligent education evaluation system provides personalized evaluation and guidance according to the individual differences of students, and promotes the personalized development of the teaching process. Fourth, high accuracy: the intelligent education evaluation system utilizes big data analysis and artificial intelligence algorithms to accurately assess the learning level of students and teaching quality [9-12]. The construction and application of intelligent teaching evaluation system in colleges and universities is an important means to improve the quality of education and teaching effect in colleges and universities. Through comprehensive and accurate assessment and analysis, the intelligent education evaluation system can provide teachers and students with accurate guidance and feedback, and help them achieve personalized learning and teaching development [13-16].

Literature [17] constructed an outcome-oriented evaluation system of teaching quality in universities with the main content of graduates’ quality, and it was found that the level of graduates’ academic and career choice competition was an important quantitative dimension of graduates’ quality. And the evaluation model of development potential index was established to provide reference for self-evaluation. Literature [18] proposed the evaluation method of classroom teaching effect of intelligent teaching mode. By constructing the intelligent teaching effect evaluation index system including teaching attitude, method and other indicators. The results verify that the proposed method improves the accuracy of the evaluation of classroom teaching effect of intelligent teaching mode and provides a new method for the evaluation of teaching effect. Literature [19] developed an evaluation method based on intelligent learning. It utilizes the restricted Boltzmann machine method based on intelligent learning to design the evaluation algorithm and evaluate the teaching quality by weighting and preference parameters. The effectiveness of this method is proved through experiments, which has the advantages of high evaluation efficiency and short delay, which is conducive to improving the teaching effect. Literature [20] developed a university classroom teaching quality information platform based on the B/S architecture model, using SSM framework and MYSQL database. Based on the current teaching status, Exibing emphasized the importance of constructing a perfect classroom teaching quality evaluation system to promote teaching quality. Literature [21] introduced the intelligent evaluation model of classroom teaching quality based on the research of classroom teaching quality evaluation index system in colleges and universities, and used the multilevel fuzzy comprehensive evaluation method to evaluate the classroom teaching quality. Example analysis specifies that the model not only can effectively evaluate the teaching quality of the classroom, but also realizes the high efficiency, network and intelligence of evaluation. Literature [22] integrates BP neural network and fuzzy mathematics theory, establishes the evaluation model of research talents in colleges and universities, and proposes the second-level index system. Based on experimental research, the effectiveness of this evaluation model is verified, and its application in practice can provide theoretical basis for related research. Literature [23] examined the design and application of FREE system combined with facial expression recognition technology based on the research of classroom teaching evaluation system. The results of the study reveal that the platform can provide decision support for classroom teaching evaluation and plays an important role in promoting the development of classroom teaching evaluation. Literature [24] describes the current status of teaching quality evaluation based on literature analysis and web survey. And the teaching quality evaluation index system was created from teaching attitude, design and other aspects. Numerical simulation was used to analyze and realize the weights of each index in the system. The simulation results illustrate that the constructed model can realize the scientific output of evaluation results and provide reference for quantitative evaluation. Literature [25] developed an online teaching quality evaluation method for basic education based on AI. The application of AI in basic education is examined and entropy weight method and gray cluster analysis method are introduced to evaluate the online teaching quality of basic education. It also proposes strategies to improve the quality of online teaching in basic education. The results of the study provide a reference for online teaching and the application of AI in basic education.

This paper mainly establishes an intelligent teaching evaluation system for higher vocational colleges based on the improved Apriori algorithm. A B/S three-layer (interface layer, business logic layer, and data access layer) is used to build the model of the intelligent teaching evaluation system. Aiming at the operational inefficiency of the traditional Apriori algorithm, the Apriori algorithm is improved by reducing the number of scanning databases to reduce the algorithmic process. In turn, the improved association rule Apriori algorithm is utilized to data mine student and teaching data to extract the needed valuable information. The potential association relationship between item sets and item sets is clarified, and the strong association rules obtained from mining are analyzed and summarized, which provides a basis for teaching decision-making in universities.

2

Design and realization of the teaching evaluation platform

This chapter first describes the logical architecture design of the system, which uses a three-tier architecture and designs the object types in the system, and then details the database design of the system, with detailed descriptions of the design of the database tables, the functional modules of the database and the evaluation.

2.1

B/S architecture

Its architecture is shown in Figure 1. B/S architecture, i.e., browser/server, is a widely used software architecture mode, which is a product of Internet technology, not only to overcome the de-electricity of the C/S architecture, but also to realize the mode of multi-user access. Using this architecture, the user only needs to access the server through the relevant browser, the server can make the appropriate response to the client’s request, presented in the client’s browser side, so the use of such architectural patterns, reducing the user side of the data processing load, but did increase the workload of the server side. Software developers only need to carry out timely maintenance, management, and upgrading of the server side. This can help maintain the system and improve efficiency of developers.

2.2

Architecture Design

This system is designed using the B/S three-tier architecture model, is the overall framework of the software is divided into three different layers, respectively, for the interface layer, business logic layer, data access layer, and each layer has a corresponding interface and components between the database to achieve the “high cohesion and low-coupling” characteristics of each layer of the specific role is described below: Data Access Layer: Operate the data information outside the database, provide effective data service and data support for the representation layer and business logic layer. Business Logic Layer: Processing the business information of the data layer, realizing the logical task operation of the data information, so as to solve the specific operation problems; Interface Layer: Used for displaying the interface information of the system or program operation, etc., realizing the interaction of human-computer information. The business between the three layers is independent and will not be affected by each other, which facilitates the developers to upgrade and maintain the different layers, and at the same time, it can also effectively improve the portability of the system and the alternative, thus reducing the development cost of the overall system design, and facilitates the system’s later maintenance and upgrading. The architecture diagram of the system design is shown in Figure 2. In the deployment of the whole teaching quality evaluation platform, the core business is the deployment of the server side, where two aspects are mainly deployed, one is the application server of the system, which mainly deploys the business logic code of the whole teaching quality evaluation platform, including the business modules of the evaluation index management, examination results management, online teaching evaluation management, teaching evaluation results and system management determined at the stage of demand analysis. Another aspect is the database server of the system, the database table in the database is responsible for storing the evaluation index information, examination result information, teaching evaluation information and some basic information in the teaching quality evaluation platform.

2.3

SQL Server database

SQL Server database is a relational database intelligent management system (RDBMS) developed and designed by Microsoft, which is now one of the mainstream databases in the world.

SQL Server has the following advantages: 1)

It has powerful features such as ease of use and scalability for distributed organizations. SQL Server databases for decision support are characterized by their close correlation with other server software and their cost-effectiveness.

2)

Data cataloging and analysis is very easy and flexible, and in a dynamic context, different departments can easily analyze data applications to stand out from the competition. This feature is significant for converting raw data into business intelligence information during data processing and parsing.

3)

The ability to update enterprise-level business applications in a short period of time greatly enhances the core strengths of an organization, giving it an unparalleled advantage over the competition.

3

Improvement of Apriori evaluation algorithm

In order to improve the quality of teaching, the school has repeatedly asked teachers to make changes in the teaching situation, so in order to fully analyze the influencing factors of the teaching situation in order to make changes accordingly, the algorithm improvement and data mining analysis are carried out with the help of data mining technology, so that the teachers can make adjustments in a timely manner according to the requirements of the quality of teaching. In the data mining analysis, for different analysis needs, often need to constantly connect with the database, thus consuming a lot of resources, and get the analysis of the data is diverse and huge, which seriously restricts the efficiency of data analysis. In order to enhance the efficiency of a large amount of data, here we analyze the two factors that have a greater impact, and put forward targeted ideas to solve the problem. On the one hand, in the data analysis will be frequent multiple connections to the database and data scanning, thus increasing the interference of invalid data, if the data can be identified, then the next scanning can be continued without having to fully scan the useless information for many times; on the other hand, the database interference with the information is more, in the data analysis is often only necessary to analyze the useful part of the data can be, but the current data analysis, the scanning of a large amount of data, the scanning of a large amount of data, the scanning of a large amount of data, the scanning of a large amount of data. The current data analysis, scanning all the data part, thus making the analysis of statistical efficiency decline, if it can be analyzed on its candidate subset, then it can meet the actual analysis needs, thus improving the efficiency of data analysis.

3.1

Definitions and Properties Related to Apripri Algorithm

Suppose the user is given the transaction database Data Base: 1 ${T_{1}, T_{2}, T_{3}, \dots, T_{n}}$

The collection Item of all individual itemsets in the transaction database is: 2 ${I_{1}, I_{2}, I_{3}, \dots, I_{m}}$

where T ⊆ Item. Association rules refer to implication formulas such as x = > y, where x is called the predecessor, y is called the successor and x and y are true subsets of Item, and the intersection of x and y must be the empty set.

Transaction support and confidence level are two key concepts in association rule mining algorithms and two evaluation metrics for determining whether an association rule is effective or not. Transaction support indicates the likelihood of an itemset appearing in all transactions of a database, while a confidence level of x = > y indicates the likelihood of itemset y in a transaction, assuming that itemset x appears in the transaction. Association rule mining searches the database for frequent itemsets that do not fall below the minimum transaction support threshold by using a minimum support threshold set by the user, and then finds rules with strong associations by maximizing the frequent itemsets.

Definition 1 The transaction support of itemset x is noted as: 3 $\sup (x)$ 4 $\sup (x) = \frac{| x |}{| D |}$

Where: $| x |$ denotes the number of times itemset x occurs in all transactions of the database and $| D |$ denotes the total number of transactions.

The confidence level of Definition 2 x ⇒ y is denoted as: 5 $c o n f (x \Rightarrow y)$ 6 $c o n f (x \Rightarrow y) = \frac{| x \cup y |}{| x |}$

Where: $| x \cup y |$ is the number of transactions that contain both x and y itemsets, and $| x |$ is the number of transactions that contain only itemset x.

Property 1 All non-empty subsets of a frequent itemset are frequent itemsets.

Property 2 All supersets of infrequent itemsets are infrequent itemsets.

Theorem 1 If the number of repeated occurrences of a single item in the set of frequent k – itemsets is less than k, then the item is not an element in the set of frequent $(k + 1)$ -itemsets.

Theorem 2 If the number of frequent itemsets in the set of frequent k – itemsets is not less than k + 1, the algorithm continues to iterate, otherwise the algorithm ends. 1

Scan the initial database and take all the individual data items that have appeared in the database as candidate 1-itemsets, and the set where the candidate 1-itemsets are located is denoted as: 7 $C_{1}$

2

Consider the candidate 1-item sets whose transaction support is above the minimum support threshold as frequent 1-item sets, and the set where all frequent 1-item sets are located is denoted as L₁.

3

For k > 1, repeat steps 4), 5), and 6).

4

Generate candidate $(k + 1)$ – itemsets from the frequent itemsets in L_k by self-joining, and the set where candidate $(k + 1)$ – itemsets are located is labeled as C_k+1.

5

Save the candidate $(k + 1)$ – itemsets whose transaction support is higher than the user-specified minimum support threshold to obtain the frequent $(k + 1)$ – itemsets, and the set where the frequent $(k + 1)$ – itemsets are located is labeled as: 8 $L_{k + 1}$

6

If the number of itemsets in L is not 0, then k = k + 1, step skip 4, otherwise the algorithm iteration ends.

3.2

Disadvantages of Apripri algorithm

The advantage of Apriori algorithm is that the idea is simple, the operation is not complicated, and the realization is easier, but the algorithm also has a lot of shortcomings, specifically, there are the following shortcomings: 1

In the process of frequent itemset self-joining to generate candidate itemsets, when there are few frequent k – itemsets, it generates many candidate $(k + 1)$ – itemsets.

2

Each candidate itemset has to scan the database once when calculating its own transaction support, and the more candidate itemsets there are, the more times the database is scanned, which makes the whole algorithm run for too long.

3

The condition of the end of the algorithm iteration is that the number of items in the maximum frequent itemset collection is empty. If there is still an item set in the frequent item set collection, the algorithm will continue until the maximum frequent item set collection is empty, which also makes the algorithm execution time increase dramatically.

3.3

Ideas for improving the algorithm

The original database is scanned and stored in an upper triangular matrix, with all individual items appearing in the database as the rows and columns of the matrix. Element A_ij of the matrix represents the probability of simultaneous occurrence of items I_i and I_j in a database transaction, and the main diagonal of the matrix represents the frequency of occurrence of a single itemset in all database transactions. Frequent 1-item sets and frequent 2-item sets are obtained from the matrix. After that, the frequent 1-item set is encoded in binary 0-1, and the length of the encoding is the number of transactions in the database, if the frequent item set exists in the transaction, it is denoted by “1”, otherwise it is denoted by “0”. Using Theorem 1, candidate k – itemset $(k \geq 3)$ is pruned once in advance, and an auxiliary table of infrequent itemsets is established, the infrequent itemsets are put into the auxiliary table, and candidate k – itemset $(k \geq 3)$ is matched with the auxiliary table, and if the match is successful, the candidate itemset is deleted, otherwise the candidate itemset is retained, and a second pruning is performed on candidate k – itemset $(k \geq 3)$ . Subsequently, traverse the frequent $(k + 1)$ – item set and perform third pruning on candidate k – item set $(k \geq 3)$ . Then using binary coding, the transaction support of candidate k – itemset $(k \geq 3)$ is obtained and finally frequent k – itemsets are obtained.

3.4

Improving the algorithmic process

1

The original database is scanned and stored in an upper triangular matrix, with all individual items appearing in the database as the rows and columns of the matrix. Element A_ij of the matrix represents the frequency of simultaneous occurrence of items I_i and I_j in the database, and the main diagonal of the matrix represents the frequency of occurrence of individual items in the database. According to the matrix, the frequent 1-item set and the frequent 2-item set are obtained.

2

The frequent 1-item set is encoded 0-1, the length of the encoding is the number of transactions in the database, if the frequent item set exists in the transaction, it is expressed as “1”, otherwise it is expressed as “0”.

3

Determine whether the number of occurrences of a single item in the frequent k – item set $(k \geq 2)$ is less than k(k ≥ 2), if it is less than $k (k \geq 2)$ , then delete the frequent k – item set $(k \geq 2)$ , otherwise keep the frequent k – item set $(k \geq 2)$ .

4

Self-connect the retained frequent k-item set $(k \geq 2)$ to generate candidate $(k + 1)$ – -item set $(k \geq 2)$ . Create an auxiliary table of infrequent item sets, put the infrequent item sets into the auxiliary table and match candidate $(k + 1)$ – item set $(k \geq 2)$ with the auxiliary table, if the match is successful, then delete this candidate item set, otherwise retain this candidate item set, and prune candidate $(k + 1)$ – item set $(k \geq 2)$ again.

5

Traverse each frequent item set in the frequent k-item set $(k \geq 2)$ , determine whether it is a subset of each candidate item set retained in C_k+1, after the completion of the traversal, remove the k-dimensional subset of the frequent k – -item set $(k + 1)$ – candidate item set, candidate $(k + 1)$ – item set $(k \geq 2)$ for the third pruning.

6

The final retention of each candidate $(k + 1)$ – itemset $(k \geq 2)$ in the use of binary coding for the phase “with”, to get each candidate $(k + 1)$ – itemset transaction support, and ultimately get the frequent $(k + 1)$ – itemset.

7

Determine whether the number of items in the frequent $(k + 1)$ – itemset is less than k + 2, if less than k + 2, the end of the algorithm, otherwise back to step 3, continue to iterate until the end of the algorithm.

4

Application of Intelligent Teaching Evaluation and Attitude of Use

This system is able to evaluate the teaching quality of colleges and universities and quickly calculate the evaluation results. Through the online network evaluation function, evaluation object scores, statistics, and analysis of evaluation object teaching quality information are collected, so as to provide objective and feasible information for colleges and teachers to improve the proposal.

4.1

Construction of the evaluation indicator system

In the process of evaluating things, the selection and design of evaluation methods is an important condition for the success of the whole evaluation function, and it is necessary to find scientific and reasonable methods of analysis and calculation. The evaluation index system is shown in Table 1.

Table 1.

Teaching quality evaluation index system

Primary indicator	Expert weight	Secondary indicator	Expert weight
Teaching attitude	0.15	Job enthusiasm	0.3
		Take care of the course	0.3
		Tutoring	0.2
		Batch job	0.2
Teaching content	0.32	Highlight	0.4
		Content enrichment	0.2
		Contact practice	0.2
		Broaden your horizons	0.2
Teaching method	0.28	Multiplicity	0.2
		Student interaction	0.3
		Teaching by aptitude	0.3
		Hands-on ability	0.2
Teaching effect	0.14	Correct opinion	0.2
		Content clarity	0.3
		Easy to understand	0.4
		Ability to improve	0.1
Teacher quality	0.11	Language	0.3
		Discipline	0.3
		Pedagogy	0.2
		Master watch	0.2

4.2

Application of Apriori algorithm in teaching evaluation

4.2.1

Data pre-processing

The data source used in this study is from the student evaluation data from the second semester of the academic year 2022-2023 in a higher education institution. Some important data from the evaluation data were selected and the data table made is shown in Table 2. There are a total of 340 teachers teaching in the school, excluding some who are on leave, leaving 323 teachers to be evaluated. After logging into the teaching evaluation system through the Android client or the web terminal, students fill in the evaluation form for each teacher, and then the system administrator exports the student evaluation data from the backend database, organizes the evaluation data, and integrates these data with the personal information of the teachers obtained from the Personnel Office, resulting in a total of 3,132 records.

Table 2.

Teaching evaluation information statistics

Teacher number	Teacher age	Teacher performance	Teaching attitude	Teaching content	Teaching method	Teaching effect	Evaluation score
0026	31	9	16	27	18	18	88
0047	30	10	15	28	17	16	86
0055	31	9	15	27	18	17	86
0133	32	10	16	26	19	18	89
0156	30	9	15	26	17	19	86
0157	32	9	17	27	18	18	89
0158	33	10	19	28	19	17	93
0159	41	9	18	29	17	16	89
0176	40	9	19	28	18	17	91
0211	32	10	18	28	16	17	89

According to the actual requirements of the algorithm and to facilitate data input, we need to preprocess the directly acquired data. According to the age level of teachers in colleges and universities, it is denoted by letter A. Teachers’ education is divided into undergraduate, postgraduate, and doctoral degrees, which is denoted by the letter B. Teachers’ titles are assistant professor, lecturer, associate professor, and professor, which are denoted by the letter F. All the basic information is discretized, and then all the indicators of the evaluation criteria (including teacher moral performance (G), teaching attitude (H), teaching content (I), teaching method (J), teaching effect (K) and total evaluation score (L), 1~5 values) in this university are discretized and converted, and combined with the exported database records, these scores are divided respectively according to the actual situation. The division method is shown in Table 3:

Table 3.

Teaching evaluation index code list

Teacher’s moral performance (10)		Teaching attitude (20 points)		Teaching (30 points)		Teaching method (20 points)		Teaching effect (20 points)		Score (100 points)
Value	Code	Value	Code	Value	Code	Value	Code	Value	Code	Value	Code
10	G1	19-20	H1	30	I1	20	J1	20	K1	95-100	L1
8-9	G2	16-18	H2	28-29	I2	18-19	J2	18-19	K2	85-94	L2
6-7	G3	13-15	H3	26-27	I3	14-17	J3	14-17	K3	81-84	L3
4-5	G4	9-12	H4	29-25	I4	9-13	J4	9-13	K4	71-80	L4
0-4	G5	0-8	H5	0-18	I5	0-8	J5	0-8	K5	60-70	L5

Discretization of the above data gives the processed data, some of which are shown in Table 4:

Table 4.

The teaching evaluation statistics of the discrete after discretization

TID	Items
T1	{A2,B1,F2,G1,H1,I3,J3,K2,L2}
T2	{A1,B2,F2,G1,H1,I2,J1,K2,L2}
T3	{A2,B1,F2,G1,H2,I2,J2,K3,L2}
T4	{A3,B1,F3,G1,H2,I1,J2,K1,L1}
T5	{A2,B1,F1,G1,H3,I3,J2,K3,L3}
T6	{A1,B3,F2,G1,H2,I2,J3,K3,L4}
T7	{A2,B1,F2,G4,H3,I4,J4,K3,L4}
T8	{A4,B2,F2,G2,H2,I2,J3,K3,L3}
T9	{A2,B3,F3,G4,H2,I2,J4,K2,L2}
T10	{A3,B2,F4,G1,H2,I3,J2,K3,L2}

4.2.2

Association Rule Mining

In order to be able to analyze the mining results more intuitively, we will combine the indicator code list and transform the codes in the above table into evaluation indicators, and the partial association rules obtained as well as their support and confidence levels are shown in Table 5, respectively. Rule 2 shows that there is a strong association rule between the title of assistant professor and the age stage between 24-35 years old, and they have a support level of 29.5% and a confidence level of 98%.

Table 5.

Mining results

Serial number	Association rule	Support	Confidence
1	I2,K2→L2	25.4%	98%
2	F1→A1	29.5%	98%
3	J2,K2→L2	27.8%	97%
4	J2,L2→K2	27.8%	92%
5	H2,L2→K2	28.5%	91%
6	J3,L3→H2	29.3%	90%
7	J2,K2→H2	25.7%	89%
8	I2,L2→K2	25.7%	89%
9	J2,L2→H2	27.7%	88%
10	K2,L2→H2	28.6%	88%
11	I3,L3→H2	27%	87%
12	I2→L2	29.3%	86%
13	A1,L3→H2	27.4%	87%
14	H2,J3→L3	29.5%	85%
15	A1,K3→J3	26%	84%
16	B2,H2→A1	29.3%	83%

4.2.3

Analysis of basic teacher information

In order to be able to follow a clear understanding of the relationship between the age and title of the teachers of this university, we visualize and analyze the age and title of the teachers of this university, and the two-dimensional scatter plot composed between them is shown in Figure 3: Through the figure, we can see that the majority of young teachers in this school, and most of the titles of the young teachers are junior or intermediate, and the titles of the teachers over 50 years of age are above the title of associate professor. According to the actual situation of the school, we can see that most of the teachers over 50 years old are retired teachers from other colleges and universities who are rehired, but relatively speaking, the proportion is small.

The analysis of the relationship between age and educational composition is shown in Figure 4, which shows the relationship between age and educational qualifications. The above figure can be analyzed to get the results: the academic qualifications of the teachers in this school include undergraduate and postgraduate students, as can be seen from the figure, most of the teachers whose age is below 35 years old have their academic qualifications as postgraduates, and the majority of undergraduates above 35 years old have their academic qualifications as undergraduates. According to the analysis of the actual situation, since the school has been established for a relatively short period of time, the situation of the majority of young teachers nowadays has emerged.

Analyzing the mining results, it can be seen: 1

We can see that there is a strong correlation between teaching methods, teaching effects, and teachers’ age and education, which confirms that teachers with relatively high teaching age and teachers with higher education are more popular with students in the teaching process, and experienced teachers are more flexible in the use of teaching methods and have a stronger mastery of knowledge, which is more recognized by the students, and therefore can also achieve better Teaching effect.

2

It is not difficult to find that there is a strong correlation between the age of teachers and their teaching attitude and evaluation scores. We can also clearly see from the scatter plot that teachers of relatively high age have higher evaluation scores, and younger teachers have relatively slightly lower evaluation results than older teachers. Therefore, in this regard, the school may consider bringing in more teachers with high academic qualifications and titles in order to improve the overall teaching level.

3

We can see that there is a strong correlation rule relationship between teachers’ qualifications, teachers’ attitudes and teachers’ age. In this, we can also see that teachers’ teaching attitudes also play an important role in the whole teaching evaluation process. Through the mining results and the above analysis, we can see that teachers’ education and title also have a certain influence on teaching quality.

4.3

Effectiveness of the use of intelligent teaching evaluation system

The wide application of “Internet+” education mode in higher vocational colleges and universities has injected new vitality into the teaching work of higher vocational colleges and universities and added strong impetus to the teaching reform of higher vocational colleges and universities. Under the guidance of intelligent education thought, the teaching of higher vocational colleges and universities is gradually moving towards intelligent teaching. At present, many higher vocational colleges and universities have adopted an intelligent teaching mode to stimulate students’ learning enthusiasm and improve the effectiveness of mathematics teaching with remarkable results. However, at this stage, there are still many problems that need to be solved. As teachers in higher vocational colleges and universities, it is necessary to change their educational concepts in a timely manner, to accumulate the greatest benefits in the educational process, and to support education and teaching with intelligence, so as to provide a guarantee for higher vocational colleges and universities to improve the effectiveness of teaching. The evaluation of the teaching effectiveness of intelligent education in higher vocational mathematics needs to comprehensively consider many aspects, such as students, teaching quality, teachers, and educational management. Through the use of intelligent evaluation tools and teaching resources in the intelligent education system, the learning effect and teaching quality of students can be improved, thus achieving objective and accurate evaluation. It is also necessary to pay attention to proper maintenance and upgrading of the intelligent education system to ensure its long-term effective application.

The intelligent platform for teaching evaluation in higher vocational colleges has been put into use for more than a year, and the effectiveness of this evaluation method can be explored by analyzing the teachers’ feeling of using it and the performance effect of the students’ acceptance of intelligent teaching evaluation. Table 6 displays the analysis of teaching and evaluation teachers’ attitudes towards the effectiveness of using the intelligent evaluation platform.

Table 6.

Analysis of platform effectiveness

Content
	1	2	3	4	5	Mean	Standard deviation	T	Sig.
Instructor (20)	0	4	4	3	9	3.85	0.742	1.357	0.194
Instructor (20)	0.00%	20.00%	20.00%	15.00%	45.00%	3.85	0.742
Evaluation teacher (35)	0	3	7	13	12	3.97	0.911
Evaluation teacher (35)	0.00%	8.57%	20.00%	37.14%	34.29%	3.97	0.911

Thirty-five evaluating teachers and 20 lecturing teachers who participated in the intelligent platform for teaching evaluation were surveyed on the effectiveness of using the platform. It was found that according to the level of agreement (1 is totally disagree, 2 is disagree, 3 is average, 4 is agree, and 5 is totally agree), more than 60% of the lecturing teachers and evaluating teachers agreed that the platform use could help them improve the quality of teaching and learning, and the proportion of the former agreeing was larger than that of the latter, but the ratings were not significantly different. This suggests that most lecturers and evaluating teachers agree that participation in teaching operations and evaluation is conducive to the improvement of higher-level teaching.

Promoting the teaching level of self and others As shown in Table 7, in terms of participation in evaluating whether or not it promotes the teaching level of self and the evaluated teachers, the percentage of lecturing teachers who agreed that it promotes the teaching level of self (80%) is much larger than that of evaluating teachers (74.29%), and there is a significant difference between the two in terms of ratings. On the other hand, the percentage of instructional faculty (75%) who agreed that they promoted the teaching standards of others was much larger than that of the evaluating faculty (34.29%), and there was a significant difference in the ratings between the two. This result suggests that participation in teaching evaluation not only enhances teachers’ own teaching level, but also helps to increase the confidence of the evaluated. What’s more, the survey shows that teaching evaluation has a greater facilitating effect on lecturing teachers, which indicates that the role of lecturing teachers’ subjectivity and initiative in participating in teaching evaluation is more evident.

Table 7.

Analysis of the effectiveness of the platform

Promote your own teaching
	1	2	3	4	5	Mean	Standard deviation	T	Sig.
Instructor (20)	1	1	2	6	10	4.15	0.652	2.345	0.022
Instructor (20)	5%	5%	10%	30%	50%	4.15	0.652	2.345
Evaluation teacher (35)	0	3	6	18	8	3.89	0.834
Evaluation teacher (35)	0	8.57%	17.14%	51.43%	22.86%	3.89	0.834

Promote the level of others’ teaching
	1	2	3	4	5	Mean	Standard deviation	T	Sig.
Instructor (20)	1	1	3	10	5	3.9	0.642	3.212	0.004
Instructor (20)	5%	5%	15%	50%	25%	3.9	0.642
Evaluation teacher (35)	0	6	12	11	6	3.49	0.914
Evaluation teacher (35)	0.00%	17.14%	34.29%	31.43%	2.86%	3.49	0.914

Overall, both evaluating teachers and lecturing teachers generally agree with the effectiveness and convenience of the intelligent platform for teaching evaluation in higher vocational colleges and universities, which is centered on the evaluation index system. This is also evidenced by the analysis of specific data on teacher development and student development (reported separately). At the same time, the teaching evaluation index system is also applicable to the teaching and activity measurement of all other courses, including art courses, with a certain degree of extensibility. Of course, due to the far from perfect design of the intelligent platform for teaching evaluation in higher vocational colleges and the complexity of teaching evaluation, this work still has the following problems and challenges:First, the system developed based on third-party applications sometimes suffers from delayed or interrupted network connections, which leads to poor use and affects the user experience;Second, unless they use evaluation indicators they are familiar with, teachers need to understand the ideas and methods of evaluation of teaching and methods in order to better utilize the platform, which is a possible difficulty in the use and promotion of the platform.

5

Conclusion

In this paper, on the basis of analyzing the operational inefficiency of traditional Apriori algorithm, an improved Apriori algorithm suitable for teaching evaluation is proposed, and an intelligent teaching evaluation system is constructed with the improved Apriori algorithm, and the system is analyzed and designed in terms of requirements. The specific empirical analysis is as follows: 1

Taking a higher vocational college as the research object, teachers with higher teaching age and higher education are more popular with students. There is a strong correlation between teachers’ age, attitude, and evaluation scores. Teachers’ education, attitude, and age have a strong correlation. Therefore, this higher education institution should introduce more teachers with higher education and higher titles to promote the improvement of teaching quality.

2

In the analysis of teachers’ attitudes toward the intelligent evaluation system, more than 60% of the teachers agree that the use of the evaluation system can help them improve the quality of teaching, and the vast majority of the teachers agree that the evaluation system can promote the improvement of their own teaching level.

Langue:: Anglais

Périodicité:: 1 fois par an
Sujets de la revue:: Sciences de la vie, Sciences de la vie, autres, Mathématiques, Mathématiques appliquées, Mathématiques générales, Physique, Physique, autres

RSS Feed de la revue

Construction and Practical Exploration of Intelligent Teaching Evaluation System in Higher Vocational Colleges and Universities

Pu Jia

Publié en ligne: 21 mars 2025

Reçu: 27 oct. 2024

Accepté: 07 févr. 2025

DOI: https://doi.org/10.2478/amns-2025-0686

Mots clésImproved Apriori, Association rules, Data mining, Teaching evaluation

© 2025 Pu Jia, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Mots clés
Improved Apriori, Association rules, Data mining, Teaching evaluation