A Bidirectional Material Diffusion Algorithm Based on Fusion Hypergraph Random Walks for Video Recommendation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper presents a strong and innovative contribution to video recommendation systems. The experimental results are also presented that show the advantages of the proposed system compared to other systems. The talked issue is interesting, but there are several concerns.

1. Lack of Clear Theoretical Justification for Model Superiority:

   • While the empirical results show improvement, the paper lacks a rigorous theoretical justification for why the HRW-BMD algorithm should outperform existing methods beyond empirical validation.
   • A deeper discussion of why bidirectional material diffusion with hypergraph-based random walks is more effective than conventional diffusion-based or bipartite graph approaches would strengthen the contribution.

2. Missing Computational Complexity Analysis:

   • The paper does not analyze the computational complexity of the proposed algorithm.
   • Given that hypergraph-based models and random walk diffusion methods can be computationally expensive, an assessment of scalability and efficiency is necessary.

3. Limited Dataset Diversity in Experiments:

   • The algorithm is only tested on the KuaiRec dataset.
   • To ensure generalizability, the model should be evaluated on additional publicly available datasets (e.g., YouTube dataset, MovieLens, Netflix dataset).

4. Insufficient Explanation of Key Parameters:

   • Several mathematical parameters used in the paper (e.g., weighting mechanisms in hypergraphs, transition probability matrices, and diffusion factors) are not adequately justified.
   • Without further explanation, reproducing the results could be difficult for other researchers.

5. No Sensitivity Analysis of Hyperparameters:

   • The paper does not explore how the choice of hyperparameters (e.g., damping factor, α, r) affects performance.
   • A thorough sensitivity analysis could help determine the robustness of the model under different settings.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Following are the suggestions to improve the quality of the submitted manuscript. 

• More clarity on dataset preprocessing and splits is needed.
• Add some recent refrences in the introduction section. 
• The paper does not thoroughly discuss the time or space complexity of the proposed model.
• While comparisons with other algorithms are thorough, ablation studies isolating the contributions of hypergraph walks, bidirectional diffusion, and negative feedback is needed.
• Discussing how this algorithm can be integrated into real-world video platforms (e.g., TikTok, YouTube) would elevate the practical impact.
• The novelty is not clear. 
• Improve the resolution of the figures. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

1. The paper proposes a Bidirectional Material Diffusion Algorithm using Hypergraph Random Walks for video recommendation, addressing sparsity and improving tag-based filtering. 2. The approach integrates hypergraph random walks and negative feedback filtering (NFDTW-TF-IDF), but some notations lack definitions, requiring a notation table and better explanations. 3. Evaluations on the KuaiRec dataset with precision, recall, and F1-score show performance improvements, but the dataset details need elaboration. 4. The study lacks a deep learning-based benchmark; adding Graph Neural Networks or Transformers would strengthen the comparison. 5. The impact of damping factor, diffusion steps, and hyperedge weights should be analyzed through an ablation study. 6. A computational complexity analysis is missing; discussing runtime efficiency would improve applicability to large-scale datasets. 7. Some grammatical issues and figure captions need revision for better clarity. Some figure captions (e.g., Figures 7 & 8) could be more descriptive. Figure3 is also not in proper format. 8. Author does not discuss biases (e.g., filter bubbles) or cold-start issues, which should be addressed. Overall Comment: 1. The study is technically sound and well-structured, but clarifications, additional baselines, and a complexity discussion would enhance its impact. 2. Addressing these issues will significantly improve the paper’s clarity, fairness, and applicability for publication readiness.

Comments on the Quality of English Language

1. The paper proposes a Bidirectional Material Diffusion Algorithm using Hypergraph Random Walks for video recommendation, addressing sparsity and improving tag-based filtering.
2. The approach integrates hypergraph random walks and negative feedback filtering (NFDTW-TF-IDF), but some notations lack definitions, requiring a notation table and better explanations.
3. Evaluations on the KuaiRec dataset with precision, recall, and F1-score show performance improvements, but the dataset details need elaboration.
4. The study lacks a deep learning-based benchmark; adding Graph Neural Networks or Transformers would strengthen the comparison.
5. The impact of damping factor, diffusion steps, and hyperedge weights should be analyzed through an ablation study.
6. A computational complexity analysis is missing; discussing runtime efficiency would improve applicability to large-scale datasets.
7. Some grammatical issues and figure captions need revision for better clarity. Some figure captions (e.g., Figures 7 & 8) could be more descriptive. Figure3 is also not in proper format.
8. Author does not discuss biases (e.g., filter bubbles) or cold-start issues, which should be addressed.
9. Overall Comment:
10. The study is technically sound and well-structured, but clarifications, additional baselines, and a complexity discussion would enhance its impact.
11. Addressing these issues will significantly improve the paper’s clarity, fairness, and applicability for publication readiness.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

All of my questions are answered in a good manner. No further comments. 

Author Response

Comments 1:[All of my questions are answered in a good manner. No further comments. ]

Response 1: [Thank you for your kind advise, your valuable suggestions are of great help to our paper.]