Knowledge Graph-Based Approach For Dynamic Ontology Generation

Tracking #: 3464-4678

Hansika Gunasekara
Thushari Silva

Responsible editor: 
Dagmar Gromann

Submission type: 
Full Paper
A knowledge graph is a data model representing real-world entities and relationships in a machine-readable format providing a comprehensive view of a specific domain or multiple domains. Dynamic ontology is a concept that refers to the idea that the fundamental nature of reality changes over time. Dynamic ontology generation using a knowledge graph automatically creates a new ontology or updates an existing ontology. This process keeps the ontology up-to-date with the changing real-world information. The most important part of dynamic ontology generation is integrating the domain ontology into the knowledge graph and updating the knowledge graph accordingly. The primary purpose of the research is to develop a dynamic environment that integrates the domain ontology with the knowledge graph. The main problem here is the dynamic integration of new data and concepts with changes in real-world information. To achieve this task graph-based ontology mapping framework is developed with matrix-based graph merging, graph clustering, cluster label propagation and matrix-based ontology mapping. The new mapping algorithm was tested with real-time dynamic data and compared with existing systems. The proposed approach outperforms the existing system in accuracy, relevancy and introducing new concepts to the ontology.
Full PDF Version: 


Solicited Reviews:
Click to Expand/Collapse
Review #1
Anonymous submitted on 14/Aug/2023
Major Revision
Review Comment:

The paper "Knowledge Graph-Based Approach For Dynamic Ontology Generation" delves into an important area of knowledge representation and offers a fresh perspective on the integration of knowledge graphs and dynamic ontology generation. However, I have identified several areas where major revisions are required to enhance the quality and clarity of the paper.

1. The abstract does not provide enough context for readers unfamiliar with the subject. It would be beneficial to briefly explain key terms such as "dynamic ontology" and "knowledge graph" and elucidate why they are significant.
2. The introduction should clearly articulate the main problem, the proposed solution, and why it is essential. Consider providing a concise problem statement early in the introduction.
3. The paper could benefit from a more detailed explanation of the new mapping algorithm. How does it compare to existing methods, and why is it superior?
4. Please elucidate the steps involved in the proposed approach, such as matrix-based graph merging, graph clustering, and cluster label propagation. Including diagrams or flowcharts might aid comprehension.
5. The section on the proposed approach (section 3) seems to lack depth. A more detailed description, perhaps with sub-sections explaining different components, would make the methodology more transparent.
6.While the literature review provides an overview of existing works, it should be more critical and analytical. It could better compare and contrast different techniques, highlighting gaps that your research fills.
7. The results section appears to lack quantitative details. It's necessary to include more specific data, statistical analyses, or comparative metrics to substantiate the claimed performance improvement.
8. Consider adding tables or charts to visually present the results. This can make it easier for readers to understand the effectiveness of the proposed method.
9. The conclusion is somewhat brief. Summarize the key findings, the implications of this work, and potential future directions in a more comprehensive manner.
10. There are several instances of awkward phrasing and minor grammatical errors throughout the paper. Please thoroughly proofread the document or consider professional editing to ensure clarity and coherence.
11. Ensure that all citations are consistent and conform to the chosen reference style. It appears that some references might be missing, especially in the introduction where you discuss the importance of ontologies and knowledge graphs.
12. If possible, consider including a case study or real-world application to demonstrate the practical relevance of your research.
13. Explain the limitations of your study and the assumptions made in the model. This provides context and helps readers understand the scope of your work.

In conclusion, the paper offers an innovative perspective on a challenging problem but requires significant revisions to meet publication standards. The revisions should focus on clarifying the methodology, enriching the literature review, providing robust and quantitative results, and improving the overall coherence and readability of the paper.

Review #2
Anonymous submitted on 05/Sep/2023
Review Comment:

This paper studies the problem of dynamic ontology generation. I think the paper can be improved in the following aspects:

1. Currently, the paper is very difficult to understand. There exist plenty of typos and format issues. For example, almost all the section titles have different cases. Some use upper cases while others use lower cases. For another example, Figures 6 and 7 are very unclear. What do you expect readers to see?

2. I don't fully understand what the authors are trying to solve. In my opinion, a knowledge graph also needs an ontology to organize its entities and relationships. So, why not directly use this ontology?

3. The paper provides several algorithms. However, they seem trivial. Furthermore, there is no complexity analysis on these algorithms.

Review #3
Anonymous submitted on 08/Oct/2023
Review Comment:

The paper deals with an interesting topic. However, I had some problems to read and understand it. For instance, the structure of Section 2 is inconsistent since at the very beginning two main classes of solutions are mentioned but, then, the structure of the section doesn't seem to actually follow that schema. Similar issues in Section 3, where some key sentences should be probably improved, as well as in most parts of the paper.

Regardless of the quality of the presentation, references should be properly extended and the dialog with literature improved accordingly. Additionally, the dataset / ontology “Onto-Travel” misses a reference, so I don’t understand if it is a pre-existing ontology or a case study defined ad hoc.

As the claim by authors ("The proposed approach outperforms the existing system in accuracy, relevancy and introducing new concepts to the ontology") is quiet ambitious, I believe it would be advisable to improve the quality of the paper and provide a more structured and consistent discussion to support such a claim.