Kochi University of Technology Academic Resource Repository

高知工科大学博士（学術）doctoral thesi

Li Chuanfa

English

Agents for DouDizhu Game Enhanced with Optimized Training Methods

香川, 拓未

高分解能磁気センサを用いた側方車両検出装置

中川, 拓海

非接触給電を用いた 小型ベアリングレスモータの開発

中澤, 巽

軸流圧縮機の翼端に加工した溝形状が失速特性と 流れ場に及ぼす影響

横内, 若菜

Japanese

微視的構造と化学組成に基づく合成ゴムの熱劣化メカニズムの解明

尾木,七海

都市集住における三種族分居モデルと相転移

遠藤, 壮真，大西, 竜司，山下, 由乃，村井, 亮介，高木, 方隆

放置竹林における竹の「資材化」と「利用拠点」の設計・施工

高知工科大学博士（工学）doctoral thesi

Lalita Na Nongkhai

－1－ 
（様式 ６） 
Form 6 
 
2025 年度 博士論文要旨 
Abstract of doctoral dissertation 
submitted in the fiscal year of  2025  
 
論文題目 Title of dissertation in Japanese (optional) 
 
 
英文題目  Title of dissertation in English 
An Adaptive Learning Support System based on Ontology of Multiple 
Programming Language 
 
学籍番号 1256010 
氏  名 Lalita Na Nongkhai 
(ローマ字表記) 
 
 
（要旨）Abstract of dissertation 
 
In recent years, programming education has become increasingly critical in developing 
students’ computational thinking and problem-solving skills. However, teaching and learning computer 
programming remains a significant challenge, particularly for novice learners. Common difficulties 
include understanding abstract concepts, mastering syntax and semantics, and applying theoretical 
knowledge to practical tasks. These challenges are compounded by the diversity of learners’ 
backgrounds and the complexity of aligning instructional content with individual learning needs. To 
address these issues, this research proposes the development of an adaptive learning support system, 
focusing on two primary studies. The first study introduced the development of ADVENTURE, an 
adaptive learning support system based on an ontology of multiple programming languages. The 
second study integrated LLMs within ADVENTURE and explored a strategy that facilitates learners 
in practicing programming to enhance their programming skills. 
In the first study (CHAPTER 2), there were two main contributions: (1) the development of an 
ontology for multiple programming languages (Java, Python, and C#), called CONTINUOUS; and (2) 
the creation of an adaptive learning support system based on this ontology, named ADVENTURE. The 
development of ADVENTURE involved two primary contributions: (1) incorporating CONTINUOUS 
into the system. Once CONTINUOUS was created, it was integrated into ADVENTURE to visualize a 
knowledge graph of programming concepts. Furthermore, CONTINUOUS is used to provide hints for 
each question and to suggest the next programming concepts to learn. These contributions demonstrate 
that an ontology can be utilized for more than just semantic representation; it can also offer significant 
－2－ 
benefits in educational contexts. For another contribution, (2) an adoption of the Elo Rating System 
(ERS) in an educational setting within ADVENTURE. Such the ERS can estimate the learners’ skills 
and exercise difficulty, it designed as adaptive mechanisms in the system for dynamically providing 
exercises that appropriate with the current learners’ skills. For experimenting in the first study 
involved 56 undergraduate IT students from a university in Thailand, which was divided into three 
groups: The preliminary group, the experimental group, and the control group. The preliminary group 
utilized an earlier ADVENTURE of version 1.0, while the experimental group utilized an updated 
version of ADVENTURE with version 1.1. The control group utilized a system of receiving random 
exercises (a mix of exercise difficulties). However, the results from learning perception show that there 
were no significant differences between all groups across four categories: mental effort, mental load, 
technology acceptance, and satisfaction. The learning logs, which recorded learning behaviors with the 
system, show that the experimental group submitted the correct answer significantly higher than both 
the preliminary and control groups (F(2,53)=5.524, p=0.007**, η²=0.172). Moreover, the rate of 
successful concepts also indicates that the experimental group had higher rates of success in completing 
programming concepts than the control group, with significant differences (p=0.031, from overall 
ANOVA results among all groups, as F(2,53)=3.032, p=0.057, η²=0.103). However, some feedback from 
participants in the first study led us to improve the next version of ADVENTURE, which we continued 
to develop in the second study. 
In the second study (Chapter 3), we explore how to enhance ADVENTURE to better support 
learners in practicing programming. This enhancement involves leveraging large language models 
(LLMs) in generative AI (GenAI) within the system. Given the rapid growth of LLMs in programming 
education, we recognize their potential to provide valuable support for learners by offering automatic 
feedback on their programming practice. Furthermore, the focus is not only on integrating LLMs 
within ADVENTURE but also on exploring which approaches aid in programming practice based on 
learners’ skills by comparing three different groups. First, Group A consisted of learners participating 
in ADVENTURE using only the adaptive mode. This updated version of ADVENTURE, now at 2.0, 
has fixed several bugs from the first study and incorporated features based on participant feedback 
from that study. Second, Group B consisted of learners participating solely in GenAI mode. This version 
of ADVENTURE focused on providing feedback and recommending exercises based only on GenAI. 
Group C consisted of learners who participated in a combination of adaptive and GenAI modes. In this 
version of ADVENTURE, automatic feedback was provided to learners, and they also had the option to 
continue to the next exercise based on recommendations from either the GenAI system or the adaptive 
mechanism. To evaluate the system, the second study was conducted with 87 undergraduate students 
from an Information Technology program at a university in Thailand. The participants, comprising 
both first year and second year students, were divided into three experimental groups: Group A 
(adaptive mode only), Group B (GenAI feedback and recommend exercise only), and Group C (combined 
adaptive and GenAI modes). The second study spanned 15 days, during which participants used the 
system to complete programming exercises. Data collection included pretests, learning logs, code 
submission records, and a post-experiment questionnaire assessing learners’ perceptions of the system. 
The results showed significant improvements in learners’ programming performance, particularly in 
the group that utilized both adaptive and GenAI modes. Learners in this group achieved higher success 
rates in submitting correct code compared to those using only GenAI, with a significant difference noted 
(p = 0.041, based on overall ANOVA results across all groups, with F(2,84) = 2.557, p = 0.084). 
Furthermore, the analysis of submitted code that lacked necessary programming logic (where learners 
attempted to guess the answers) indicated that Group B had the lowest instances of submissions 
without required programming logic compared to Group A (p < 0.001) and Group C (p = 0.001), 
demonstrating significant differences among all groups (F(2,84) = 8.185, p < 0.001**). Additionally, 
other results show that Group C have greater results compared to both Group A and B. 
－3－ 
In conclusion, this study presents the design, development, and evaluation of an innovative learning 
support system called ADVENTURE. This system integrates ontology-based knowledge 
representation, adaptive learning mechanisms, and feedback generated by large language models 
(LLMs). ADVENTURE provides a scalable and effective solution to the challenges faced in 
programming education by offering personalized, data-driven, and conceptually grounded instruction to 
learners. This research establishes a foundation for the next generation of intelligent educational tools, 
combining the strengths of semantic web technologies and generative AI to meet diverse learning 
needs.  
 
Keywords: adaptive learning; ontology; large language models; programming exercise; recommending 
exercises; multiple programming language 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
（注）論文要旨：内容は4,000字程度とし、電子ファイル及び紙媒体で提出 
The abstract of doctoral dissertation should be approximately 2,000 words and must be submitted in both 
electronic and printed formats. 


An Adaptive Learning Support System based on Ontology of Multiple Programming Language

中道, 誠

気候変動による高潮・高波への影響評価と海浜変形の分析手法に関する研究

竹又, 裕也

アルミニウム合金/CFRP重ね合わせ接着継手の 破壊挙動に及ぼす被着体剛性および接着条件の影響

Kochi University of Technology

Not a member yet

Agents for DouDizhu Game Enhanced with Optimized Training Methods

高分解能磁気センサを用いた側方車両検出装置

非接触給電を用いた小型ベアリングレスモータの開発

軸流圧縮機の翼端に加工した溝形状が失速特性と流れ場に及ぼす影響

微視的構造と化学組成に基づく合成ゴムの熱劣化メカニズムの解明

都市集住における三種族分居モデルと相転移

放置竹林における竹の「資材化」と「利用拠点」の設計・施工

An Adaptive Learning Support System based on Ontology of Multiple Programming Language

気候変動による高潮・高波への影響評価と海浜変形の分析手法に関する研究

アルミニウム合金/CFRP重ね合わせ接着継手の破壊挙動に及ぼす被着体剛性および接着条件の影響

2,359

2,367