Analysis of AI Immersive Interpretation Teaching Evaluation Based

on Data Mining

Siying He

Foreign Language Department, Sichuan University Jinjiang College, Meishan, Sichuan, 62080, China

Keywords: VR, AR, Immersive Teaching, Embodied Cognition, Data Mining.

Abstract: The rapid development of China has led to an urgent demand for qualified English interpreters. However,

the interpretation classes in universities fail to build immersive teaching spaces as well as corresponding

teaching resources and evaluation systems, which makes it difficult to meet the demand. Upon the trend of

intelligentization of foreign language education. This study introduces the operating mechanism and

characteristics of AR and VR, constructs an AI immersive interpretation teaching evaluation model, and

uses data mining technology to mine and analyze teaching-related data. Based on the theory of embodied

and immersive teaching, five- dimensional evaluation index is established to accurately reflect the learning

effect of students. This paper focuses on the application of data mining technology in AI immersive

interpretation teaching evaluation, which can promote the quality of school teaching and has great

significance for AI immersive interpretation teaching evaluation.

https://orcid.org/0000-0002-9078-6064

1 INTRODUCTION

With the rapid development of emerging

technologies such as AI, 5G, virtual reality (VR),

internet of things and big data, human society has

entered an era of intelligence. Among them, the

massive use of extended reality (XR) and

holographic technology is creating lifelike virtual

environments, providing people with new immersive

experience. “The Action Plan of Education

Informatization 2.0” by Ministry of Education

(2018) clearly proposes to “carry out the reform of

education mode with new technology support and

accelerate the construction of intelligent learning

spaces with virtual simulation experimental teaching

projects”. Upon this policy, AI technologies have

led to significant changes in language teaching

industry. Meanwhile, with the implementation of

such new national initiatives as “One Belt, One

Road”, “Building a Shared Future of Human Being”

and “Telling the Chinese Story”, China is in an

urgent need of qualified interpreters with specialized

abilities and cross-cultural awareness. In “Teaching

Guidelines for Undergraduate Foreign Language and

Literature Majors in General Higher Education

Institutions”, Ministry of Education (2020) proposed

that modern technologies should be integrated into

interpretation teaching. Thus, it can be seen that

actively introducing AI technologies to improve or

innovate teaching mode is an effective way to

reconstruct interpretation education ecology and

successfully cultivate high-quality interpreters for

foreign trade and cultural communication of China.

Data mining is mainly cloud architecture

computing technology, and cloud architecture

computing is the development and integration of

distributed computing, Internet technology and

large-scale resource management technology. Its

applications and research include resource

virtualization, information security, and massive

data processing (Luo, and Li, 2014). The structure of

the data mining system is shown in Table 1.

192

He, S.

Analysis of AI Immersive Interpretation Teaching Evaluation Based on Data Mining.

DOI: 10.5220/0011908900003613

In Proceedings of the 2nd International Conference on New Media Development and Modernized Education (NMDME 2022), pages 192-196

ISBN: 978-989-758-630-9

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

Table 1 Structural Diagram of Data Mining System

2 AI TECHNOLOGIES

APPLICABLE IN IMMERSIVE

TEACHING OF

INTERPRETATION

Based on the embodiment-based immersive teaching

theory, we choose the following AI technologies to

empower the practice of the theory.

VR, AR and other XR technology are the core

for creating immersive teaching environments. VR

technology mainly includes dynamic environment

modeling, 3D graphics generation. Based on

computer technology and various display devices,

combined with generation, multi-sensing interaction

and display technologies, VR technology allows

users to interact with each other and the

environment in 3D space in real time. AR

technology, on the other hand, mainly includes 3D

registration, human-computer interaction. AR

technology provides users with holographic glasses

and other devices to perceive information and the

environment. Mixed Reality technology is a fusion

of VR and AR.

All three work logically by first inputting audio

and image signals and then using computers,

smartphones and other devices to present a

simulated environment. Then, through voice

recognition, voice interaction, and human-computer

dialogue, users can interact with the virtual world

and reality through the computer. They construct a

“real” learning context, and enhance the “sense of

presence” of learners. In general, XR technologies

provide experience and contextual support for

immersive teaching.

Table 2 The Working Mechanism for VR, AR and XR

3 CONSTRUCTION OF AN

EVALUATION MODEL FOR AI

IMMERSION

INTERPRETATION TEACHING

3.1 Establishment of Five Dimensions

Evaluation Index

The curriculum of “AI immersive interpretation

teaching” is selected as the action research sample.

This study combines process evaluation and

summative evaluation and constructs a

five-dimensional evaluation model for blended

learning from three aspects: learning process,

learning effect and learning attitude. Among them,

the learning process evaluation mainly investigates

the students’ participation and interaction of online

courses; The evaluation of learning results is

realized by the effectiveness degree. The evaluation

of learning attitude was calculated by the

questionnaire of adaptability and satisfaction.

(1) Participation

As an example, students must participate in the

entire course from task design to implementation

(Song, and Chen, 2018). In this study, only students’

online participation was calculated, which mainly

includes login times (Q1), times of entering courses

(Q2), times of submitting course assignments (Q3),

times of participating in course questionnaires (Q4)

and online hours (Q5).

Analysis of AI Immersive Interpretation Teaching Evaluation Based on Data Mining

193

For ease of calculation, this study uses a

five-point Likert scale to calculate each dimension

of the evaluation model. Assume that participation

Q=Q1+Q2+Q3+Q4+Q5, each of which counts for 1

point to simplify the calculation；Assuming that the

actual value of the above items is Qi (i=1,2,3,4,5),

the theoretical value is Q'i, n is the number of

parameters, and the total number of people is N,

then the participation can be calculated as follows:

Q=





×N



n=5







（1）

Qi/N is the actual average value of a parameter,

and Qi/(N×Q' i) is the actual value/theoretical value,

that is, the degree of completion of the theoretical

value. If Qi/(Q' i×N)≥1, the value of this item is 1,

and the theoretical goal is achieved ； If Qi/(Q'

i×N)<1, the theoretical goal is not achieved. The

value of n will also increase with the iteration of

THEOL platform, and the formula (1) from this

study can also be applied.

(2) Interaction

Due to the unavailability of data from real-time

interactive classroom, but only network interaction,

this study mainly include the number of reading

teaching materials (J1), the number of class

discussion topic (J2), group discussion area (J3),

according to the number of palindrome course

discussion is palindrome number (J4) and the

number of reading course notice (J5).

J=





×N



n=5







(2)

Ji/S is the actual average value of a parameter,

and Ji/(N×D 'i) is the actual value/theoretical value,

that is, the degree of completion of the theoretical

value. If Ji/(J 'i ×N)≥1, then the value is 1, and the

theoretical goal is achieved. If Ji/(J' i×N)<1, the

theoretical goal is not achieved.

(3) Adaptability

Since students have different levels of

knowledge and experience, the adaptability of

blended learning will be different. The adaptability

of blended learning mainly examines students’

recognition of the new learning style (R1), task

difficulty (R2), teamwork (R3), re-engagement (R4)

and recommendation index (R5).

R=





×N



n=5







（3）

Ri/N is the actual average value of a parameter,

and Ri/ (N×R' i) is the actual value/theoretical value,

that is, the degree of completion of the theoretical

value. If Ri/(R 'i ×N)≥1, then the value is 1, which

achieves the theoretical goal. If Ri/(R' i×N)<1, the

theoretical goal is not achieved.

(4) Satisfaction

In distance training, student satisfaction is

usually used to evaluate student response. Trainees’

response to the training theme, online training

course, online tutor, and online training organization

make up the main aspects of online training.

(5) Effect

This study proposes that the effectiveness

evaluation of students’ learning effects of blended

learning mainly includes two parts: ① the

evaluation of students' experimental results is carried

out by combining teachers' evaluation and students'

mutual evaluation; ② The students' mastery of

thematic skills was investigated by questionnaire.

(Yin, and Qiao, 2017)

3.2 SVM Algorithm for Evaluating

Teaching

A Quadratic programming will be used to solve the

support vector machine (SVM) algorithm. (Zhang,

Li , and Fu, et al. 2005) In the case of the existing

samples, assuming that ωx + b=0 on the sample

set defines two classes. Which training

sets





,yi







∈R，y



C−



−1,+1



i=1,⋯n then

have a hyperplane can make two types of samples

and its distance can reach the maximum, then the

plane as the hyperplane, and hyperplane calculation

formula is:

‖



ξ







ξ≥0,i=1, ,2,n



（4）





=sgnα

















+b

（5）

To sum up, this study evaluated students’

blended learning in online courses from five

dimensions, including participation, interaction,

effect, adaptability and satisfaction, and expected to

improve the blended teaching design through the

NMDME 2022 - The International Conference on New Media Development and Modernized Education

194

evaluation results, so as to provide a basis for the

next teaching practice. (Zhao, Shi, and Wang, 2006)

4 AI INTERPRETATION

TEACHING QUALITY

EVALUATION DATA ANALYSIS

AND DISCUSSION

Students’ participation, satisfaction, and effect

degree of online courses are relatively high based on

the results of the five-dimensional evaluation model

(Table 4)；One semester of study showed good

adaptation to blended learning, but poor interaction

with it. The results also show that hybrid learning

has achieved obvious results. (Romdhani S, Torr P,

Scholkopf B, et al. 2004). In particular, it has

achieved remarkable results in enhancing students

participation enthusiasm, strengthening operation

consciousness and exercising knowledge and skills.

In the classroom, the simulated environment

integrates various factors of context, text and

situation in the interpretation through VR and AR.

In traditional interpretation training, students only

need to complete the basic task of language

conversion. In the virtual classroom, students have

to face all kinds of challenges in the virtual world,

and accidents may happen at any time. Using the

five-dimensional evaluation model, the interaction

of students on the network platform is mediocre. As

a result, not only is there no student interaction on

the network platform, but the hybrid teaching

practice also suffers from poor interaction design.

Table 3 Learning Evaluation Results

Projec

Participatio

Interactio

Adaptabilit

Averag

A1 3 5 3 3.7

B2 4 4 3 3.7

C3 3 4 4 3.7

D4 4 4 3 3.7

E5 5 4 3 4.0

F6 4 4 4 4.0

G7 5 2 4 3.7

H8 4 3 5 4.0

I9 3 4 2 3.0

J1 4 3 4 3.7

Table 4 Learning Evaluation Results of the Evaluation

Model

5 CONCLUSION

The immersive learning environment provides

interpreters with visually, audibly and emotionally

interactive experience through AI technologies. In

other words, it is an environment where the

physical, the social, the cultural, and the

psychological situation are simulated. It enables

learners to input and input language scientifically

and efficiently with the support of the environment

so as to realize the internalization of language

ability. More importantly, in the environment, an

interpreter’s meta-cognitive level can be enhanced,

which in turn promotes the role of their

non-intellectual factors and enhances their learning

effectiveness. On the other hand, the dynamic and

open nature of the immersive embodied teaching

environment makes the time and space of teaching

more diversified and the relationship between

teachers and students more equal.

This study expects to improve the

five-dimensional evaluation model of mixed

learning through teaching practice. Based on three

aspects: learning attitude, learning process, and

learning effect, this study constructs a

five-dimensional evaluation model for blended

learning. In this study, the blended learning model is

Analysis of AI Immersive Interpretation Teaching Evaluation Based on Data Mining

195

used as a framework to guide immersive

interpretation teaching. Finally, through data

collection and questionnaire survey, students were

evaluated from five dimensions of participation,

interactivity, adaptability, satisfaction and

effectiveness, and the design of blended learning

was constantly fed back and adjusted. In addition to

improving the learning effect of students, the

five-dimensional evaluation model verifies the

effectiveness of the model and its positive incentive

effect.

REFERENCES

Chen Guang-xi, Xu Jian, Cheng Yan. (2009). Cluster

method of support vector machine to solve large-scale

data set classification. J. Computer Science.3,

184-188.

Luo Xiao, Li Shan. (2014). Construction of research-based

Teaching evaluation System for Statistics Major

Curriculum. J. Curriculum Education Research. 7,

240-241.

Ministry of Education. (2018). Education

Informatization 2.0 Action Plan. [online] available at

<http://www.moe.gov.cn/srcsite/A16/s3342/201804/t2

0180425_334188.html>[Accessed 18 April 2018].

Ministry of Education. (2020). Teaching Guide for

Undergraduate Foreign Language and Literature

Majors in Ordinary Colleges and Universities (I) -

Teaching Guide for English Majors. N. Foreign

Language Teaching and Research Press.

Romdhani S, Torr P, Scholkopf B, et al. (2004). Efficient

face detection by a cascaded support-vector machine

expansion. J. Pattern Recognition. 62-70.

Song Cheng, Chen Xiao. (2018). Empirical Research on

the Learning Effect of online Open Courses for Higher

Vocational Students. J. Shaanxi Education (Higher

Education).11, 5-56+69.

Yin Yang, Qiao Ailing. (2017). Analysis and Design of

Flipped Classroom Learning Evaluation based on

MOOC platform. J. Information Technology

Education in primary and secondary schools. 5, 38-42.

Zhang Min, Li Wuping, Fu Jufang, et al. (2005).

Application of clinical nursing teaching quality

evaluation database in clinical practice management.

J. Nursing Research.19, 1773-1774.

Zhao Xiu-feng, Shi Ke-Xue, Wang Chuan-long. (2006).

Design and Implementation of College Teaching

Quality Evaluation System Based on ASP and Web

Database. J. Educational Information Technology. 13,

54-55.

NMDME 2022 - The International Conference on New Media Development and Modernized Education

196