FUNCTIONS AND APPLICATION EFFECTS OF COURSE KNOWLEDGE GRAPHS - A CASE STUDY OF WESTERN CIVILIZATION HISTORY

[Fund] Provincial College Student Innovation Training Program of Guangzhou University

[Фонд]Программа инновационного обучения студентов Университета Гуанчжоу, Китай, № S202411078127

ABSTRACT

Course knowledge graphs serve as a critical tool for the digital transformation of higher education. This paper elaborates on the core functionalities of knowledge graphs, encompassing multi-dimensional connections of knowledge points, integration and updating of multi-modal resources, personalized learning support, and real-time evaluation and feedback. Taking the Western Civilization History course as an example, it discusses the significance of knowledge graphs in reconstructing teaching content, enabling personalized learning paths, and effectively achieving instructional objectives. Empirical analysis demonstrates the role of multi-modal resource-enhanced knowledge graphs in enhancing learning outcomes, providing practical insights for educators to develop and apply course knowledge graphs.

АННОТАЦИЯ

Граф знаний учебного курса является важным инструментом цифровой трансформации учебных курсов в высшем образовании. На основе описания функций графа знаний, таких как многомерная связь ключевых понятий, интеграция и обновление мультимодальных ресурсов, обеспечение персонализированного обучения, а также оперативная оценка и обратная связь, в данной статье на примере графа знаний курса по истории западной цивилизации анализируется его значение для реконструкции учебного содержания, формирования индивидуальных образовательных траекторий студентов и эффективного достижения педагогических целей. Эмпирическое исследование демонстрирует влияние графа знаний на повышение успеваемости, что предоставляет практический опыт для преподавателей по разработке и применению графов знаний курса.

Keywords: Course, Knowledge Graph, Function, Effectiveness, Western Civilization History.

Ключевые слова: курс, граф знаний, функции, эффекты, история западной цивилизации.

Introduction

The digital transformation of higher education serves as a critical pathway for constructing a modern, powerful education system and an essential means for cultivating talent to meet the demands of modernization. In March 2022, the launch of the smart education platform of China marked the beginning of a strategic initiative to vigorously advance educational digitalization across the country[1]. To facilitate the digital transformation of education, Chinese universities have actively participated in the construction and research of course knowledge graphs.

Currently, the achievements of Chinese academic circles on course knowledge graphs can be primarily categorized into four types. First, studies on the construction technologies of course knowledge graphs. Much of this research is based on knowledge graphs independently developed by faculty in science and engineering disciplines. For example, Zhan Wen et al. investigated the construction technologies of knowledge graphs and evaluation modules, using the Information Theory and Coding course as a case study[2,p.6]. Second, research on reforming courses or course systems through the application of knowledge graphs. There is a substantial body of work in this area. For instance, Guo Caili et al. explored how knowledge graphs can empower the construction of course systems[3,p.54]. Third, the construction and application research of knowledge graphs on online teaching platforms. These studies are often conducted by frontline educators focusing on specific courses. Examples include Chen Yu et al., who examined the construction and application of knowledge graphs under university-industry collaboration[4,p.119], and He Yu et al., who studied the development and application of knowledge graphs for university physics courses[5,p.149]. Fourth, comprehensive research on the development, application, and effectiveness of course knowledge graphs. For example, Qu Kechen et al. studied the design techniques and usage outcomes of learning systems based on knowledge graphs[6,p.70].

Moreover, there is still a lack of empirical research on the functionalities of knowledge graphs and their effectiveness in enhancing student learning outcomes. Based on the experience of constructing a knowledge graph for the Western Civilization History course, this paper explores the potential functions of knowledge graphs and provides empirical evidence to demonstrate their role in improving student learning performance.

1. Functionalities of Course Knowledge Graphs

The functionalities of knowledge graphs can be summarized into four aspects: enabling multi-dimensional connections of knowledge points, integrating multimodal resources, supporting dynamic updates of resources, and facilitating personalized autonomous learning while generating learner profiles.

1.1 Creating multi-dimensional connections of knowledge points

Knowledge graphs possess the functionality of establishing relationships between knowledge points. By configuring these relationships, teachers can define a learning sequence for the knowledge points. Students can either follow the teacher-defined sequence or customize their learning path by selecting the content and order. The common relationships between knowledge points include inclusion, sequential, and relevant relationships, along with their sub-relationships.

Inclusion relationship. It refers to a situation where one object or concept encompasses another object or concept, or where a larger entity contains one or more smaller entities. For instance, the knowledge point "Signs of Civilization" includes the knowledge point "Writing".

Sequential relationship. It can be further refined into sub-relationships such as “Sequence”, “Dependency”, and “Extension”.

Through the "Sequence" relationship, educators can establish a primary logical thread in the knowledge space that aligns with the teaching order. For instance, after studying the knowledge point "Ancient Aegean Civilization," students can proceed to the next topic, "Ancient Greek Civilization," following the sequential logic prescribed by the educator.

Through the "Dependency" relationship, it is determined that the formation of the previous knowledge point depends on the latter knowledge point, and the latter knowledge point is the condition for learning the previous knowledge point, for example, the knowledge point "Civilization" depends on the knowledge point "Signs of Civilization".

Through the “Extension” relationship, it assists to extend the thinking and learning of other knowledge from one knowledge point. For example, from the knowledge point “Peloponnesian War,” it extends to the study and thinking of “Thucydides' Trap Theory” in the field of international relations. The extension relationship can assist teachers associate relevant knowledge points in different disciplines, forming an interdisciplinary knowledge network. It is an important means to break down disciplinary barriers and reconstruct teaching content.

Relevant relationship. It can be refined into sub-relationships such as “Comparative,” “Symbiosis,” and “Similarity.”

The “Comparative” relationship refers to the examination of similarities and differences between two or more entities in specific aspects. By establishing “Comparative” relationships, educators can assist students in better understanding the similarities and differences between knowledge points or concepts, thereby fostering a more comprehensive grasp of a topic or problem and providing additional analytical perspectives.

The “Symbiotic” relationship refers to two knowledge points that both address different aspects of the same theme or share common attributes. By establishing a "Symbiotic" relationship, educators can provide a comprehensive overview of a subject. For instance, setting the knowledge points "Athens" and "Sparta" as a symbiotic relationship allows for a thorough introduction to the history of city-state development during the Classical Period of Greece.

The “Similarity” relationship refers to the connection between knowledge points that share identical or comparable characteristics, attributes, or functions. By setting a "Similarity" relationship, educators can link related knowledge points together.

In the knowledge graph, various complex relationship types such as causality, derivation, and application can also be set between knowledge points. A knowledge point can have multiple relationships at the same time, and these relationships are intertwined, forming a multidimensional connection network. For example, the “Peloponnesian War” not only incorporates similarity or comparative relationships with other wars but also incorporates an extension relationship with “Thucydides' Trap Theory.” This multidimensional connection enables knowledge points to form a comprehensive and three-dimensional network, providing conditions for personalized learning.

1.2 Mounting and Updating Multimodal Resources

In the knowledge graph, multimodal resources such as textbooks, courseware, pictures, videos, texts, links, and test questions are mounted, which can present knowledge from multiple dimensions.

Teacher-Led Resource Mounting and Updating

Teachers can mount and optimize multimodal resources in the knowledge graph based on instructional objectives and students' actual learning progress. They can also incorporate the latest academic findings and pedagogical concepts into resource updates, introducing cutting-edge knowledge to the knowledge graph and broadening students' knowledge horizons. Additionally, teachers can create exercises tailored to specific knowledge points and associate them with the corresponding knowledge points in the question bank. Once the association is completed, the system will automatically select and present relevant exercises to students each time they access the knowledge point for practice.

Platform-Driven Automatic Mounting and Updating

The platform possesses extensive resource reserves and technical capabilities, enabling the integration and large-scale updating of multimodal resources in the knowledge graph from a macro perspective. Platform administrators can classify and organize scattered resources according to criteria such as subject areas and knowledge points, and then mount them. For example, resources such as images and videos uploaded by different teachers for the same subject-specific knowledge point can be integrated, uniformly annotated, and mounted to the corresponding knowledge nodes in the knowledge graph, thereby enhancing the standardization and usability of the resources.

AI-Driven Recommendation and Updating

AI can update multimodal resources in the knowledge graph based on factors such as each learner's progress and preferences, providing personalized learning solutions. By analyzing data such as learners' performance on exercises and study duration, AI can identify knowledge gaps and recommend updated resources, such as textbook content and test questions, to strengthen learners' mastery of these weak areas. AI not only assists teachers enhance instructional effectiveness but also alleviates teaching pressure, significantly improving both teaching and learning support functions.

1.3 Enabling Personalized Learning

The course knowledge graph provides a robust foundation for personalized learning. It not only leverages AI assistants to recommend targeted learning resources based on students' performance but also constructs learner profiles. By analyzing students' responses to exercises, the system can recommend relevant internal resources, external resources, and test questions tailored to individual requires.

Self-Directed Learning

In knowledge graph-assisted learning, students can independently explore knowledge. For example, in the knowledge graph of a Western Civilization course, students begin with the democratic system of ancient Greece and, through the knowledge graph's associations, connect to related concepts such as the Roman Republic, thereby constructing a personalized knowledge system rather than following a traditional linear sequence. Additionally, students can control the pace of their learning based on their abilities and time availability.

Learner Profiling

The knowledge graph analyzes the knowledge students have mastered and the areas requiring improvement, integrating factors such as their cognitive levels, interest preferences, and learning histories to assess their learning status and construct learner profiles. Based on these profiles, the system intelligently recommends personalized learning paths to guide students. By leveraging these profiles, the knowledge graph enables precise resource pushing and dynamically adjusts learning paths based on students' real-time feedback and progress, ensuring that the recommended content consistently aligns with their individualized learning requires.

1.4 Real-time evaluation and feedback

The course knowledge graph incorporates real-time evaluation and feedback functions.

Real-time evaluation.

Teachers evaluate student learning in real-time by embedding tasks such as quizzes, assignments, or exams in the graph.

Real-time feedback.

After students complete practice questions or tasks, the knowledge graph system can immediately evaluate their learning effects, such as correct rate, answering speed, etc., and feedback the evaluation results to students and teachers. This kind of immediate feedback assists students to keep abreast of their own learning and adjust their learning strategies, and it also assists teachers to keep abreast of students' learning dynamics and intervene in a timely manner.

2. Analysis of the application effect of the course knowledge graph

In order to test whether the knowledge graph of the course can improve the learning effect of students, the course of History of Western Civilization was experimented in two parallel classes in two rounds of teaching practice and collected feedback from students on the utilize of knowledge graphs.

The initial round of the teaching experiment was conducted in April 2024, with 26 students participating in each class. The primary objective was to evaluate the effectiveness of utilizing the course knowledge graph, with the experimental content focusing on the knowledge point of the “Age of Discovery”.

In the first round of teaching utilizing the course knowledge graph, the "Age of Discovery" knowledge point was selected for experimentation. First, a pre-test with identical content was administered to two classes. The pre-test scores of the two classes were then analyzed using the non-parametric Wilcoxon rank-sum test. As displaed in Table 1, the significance level (p-value) between the two groups was 0.122, which is higher than the threshold of 0.05, indicating no statistically significant difference in performance between the two classes.

Table 1.

Wilcoxon Rank-Sum Test Results for Pre-Test Scores on the “Age of Discovery” Knowledge Point

Then, the traditional textbook preview task was assigned to the control class, and the knowledge graph knowledge point preview task with text learning materials was assigned to the experimental class. After the students completed the task, the two classes were assigned post-test with the same content, and the non-parametric Wilcoxon rank sum test was performed on the post-test scores of the two classes, as displayed in Table 2, the significance level value of the two components distribution was 0.007, which was lower than the significance level of 0.05, indicating that there was a significant difference in the scores of the two classes.

Table 2.

Wilcoxon Rank-Sum Test Results for Post-Test Scores on the “Age of Discovery” Knowledge Point

Combined with the fact that the average score of the experimental class (86.4) was higher than that of the control class (76.6), the experimental results display that the knowledge graph of mounting text resources can improve the learning effect.

At the conclusion of the first round of teaching, students completed a survey. The survey results indicated that the knowledge graph was beneficial in assisting them to clarify relationships between knowledge points, master the content, expand their knowledge base, and enrich their knowledge system. Additionally, it facilitated interdisciplinary learning and research, contributing to the enhancement of their comprehensive qualities and innovative capabilities. However, students expressed a desire for the inclusion of more multimodal materials, such as videos, to further improve the effectiveness of the knowledge graph.

Based on the feedback from students, instructors mounted various multimodal resources, including images, textbooks, academic literature, and micro-lecture videos, into the course knowledge graph.

The second round of the teaching experiment was conducted in October 2024, with 25 students participating in each class. The primary objective was to evaluate the effectiveness of the updated course knowledge graph, with the experimental content focusing on the knowledge point of “Cities of Medieval European”.

During the experiment, a pre-test was administered to both classes. The pre-test scores were analyzed using the non-parametric Wilcoxon rank-sum test, and the results displayed no statistically significant difference between the two groups. Subsequently, the control class was assigned traditional textbook-based learning tasks, while the experimental class was assigned tasks utilizing the knowledge graph for learning the knowledge points. As illustrated in Figure 1, the knowledge graph displayed the knowledge point through conceptual hierarchical relationships and mounted multimodal learning resources for multiple concepts.

Figure 1. Knowledge Graph of “Cities of Medieval European” Knowledge Point

The students completed the test of the same content after completing the task. The results of the nonparametric Wilcoxon rank sum test are displayed in Table 3.

Table 3.

Wilcoxon Rank-Sum Test Results for Post-Test Scores on the European Cities Knowledge Point

The results in Table 3 display that the value of the significant level of achievement between the two groups is 0.002, which is lower than the significant level of 0.05, indicating that there is a very significant difference in the performance of the two groups. Combined with the fact that the average score of the experimental class (87) is higher than that of the control class (73.3). The experimental results demonstrate that the course knowledge graph, enriched with multimodal resources, effectively enhances learning outcomes.

By comparing the two rounds of teaching practices, it can be concluded that supplementing and developing multimodal resources within the knowledge graph contributes to improving students' academic performance. However, the functionality of the knowledge graph still requires further development, particularly in constructing personalized learning pathways. Additionally, more extensive empirical research is required to validate the long-term impact of knowledge graphs on student learning. 

3. Conclusion and suggestion

The course knowledge graph incorporates a variety of functionalities, and teachers can construct the course knowledge graph on the teaching platform by themselves or in cooperation with professional teams. Teachers and students can utilize the knowledge graph to teach and learn before, during and after class, and the experimental results display that the knowledge graph mounted with multimodal resources can effectively improve students' academic performance.

At present, the construction of course knowledge graph is still in the development stage. Whether it is the teaching practice of front-line teachers or the technical support of the platform, its function development and application mode are still in the value verification period. Driven by the iterative upgrading of artificial intelligence technology and the urgent require for precise talent training, knowledge graph technology will accelerate the continuous expansion of functions. Relying on the ever-evolving knowledge graph tools, teachers can systematically promote the digital transformation process of the course system, thereby providing strong support for the construction of a modern education power.

References:

1. Steadfastly Advancing the Digitalization of Education in China / [EB/OL]. - Access mode: http://www.moe.gov.cn/jyb_xwfb/s5148/202403/t20240329_1122956.html. (accessed: 24.12.20).
2. ZHAN Wen, HU Zhe, JIANG Yuan, SUN Xinghua. Construction of a Formative Curriculum Evaluation System Based on Knowledge Graphs: A Case Study of Information Theory and Coding Courses// Survey of Education. 2024(31) [in Chinese]
3. GUO Caili, LIU Fangfang , YANG Yang. Exploration and Research on Empowering Curriculum System Construction with Knowledge Graphs: A Case Study of Signal and Information Processing Courses// China University Teaching. 2024(11) [in Chinese]
4. CHEN Yu, LIU Zhi-ming, WANG Bin-jie. Exploration of Knowledge Graph Construction under the Online Internet Course Platform: Taking Rail Vehicle Equipment as an Example// Education and Teaching Forum. 2024(49) [in Chinese]
5. HE Yu, SUN Yanyun, XIE Dong et al. Construction and practice of college physics curriculum based on knowledge graph// Physics and Engineering. 2024(06) [in Chinese]
6. QU Kechen, LI Jinchan, HUANG Deming et al. Study on the influence of a knowledge graph-based learning system design on online learning results// Journal of East China Normal University(Natural Science). 2024(05) [in Chinese]