The Recommendation Algorithm Based on Improved Conditional Variational Autoencoder and Constrained Probabilistic Matrix Factorization

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. There are a number of typos and formatting issues, thorough proof read is recommended.
2. Authors have validated the proposed algorithm considering the use case of MovieLens 100K dataset. One example may not be sufficient to prove the efficiency of the algorithm. Authors must consider other applications to further elaborate the effectiveness of the proposed algorithm.
3. It is recommended to present the algorithm using the pseudocode for better understanding.
4. A flow chart may help to improve the readability of the article.
5. Compare the proposed algorithm computational and space complexity with other other algorithms considered in the paper, comparison with other algorithms is also highly recommended to prove the efficiency of the proposed algorithm.
6. Algorithm flow must be presented with graphical contents.
7. The resolution of all the figures must be improved and axis must be properly explained in the text. Some terms such as epochs and implicit factor has been used in figures but clearly explained in algorithm. Why the term epoch has been used for iterations?

Comments on the Quality of English Language

It seems to be OK.

Author Response

Dear reviewer
Thank you very much for your comments and suggestions.
Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches.   We have studied comments carefully and have made correction which we hope meet with approval.   The modifications are explained in the paper in the form of notes.
The following is a reply to your suggestion:
Suggestions 1:
The format of the full text has been checked and the format problem of formula numbering has been modified, which will continue to be refined after the review of subsequent editors.
Suggestions 2:
The MovieLens-1M dataset is added in this paper, and supplementary experiments are carried out in the fourth chapter of the paper, and the experimental results are explained.
Suggestions 3,4 and 6:
Thank you very much for your suggestion, pseudo-code and flowchart can indeed help to better understand the algorithm, so the flowchart is selected between flowchart and pseudo-code, and the flowchart is added in Section 3 of the paper and marked.
Suggestions 5:
Increasing computational complexity and time complexity can indeed better reflect the efficiency of the proposed algorithm. However, RMSE and MAE were selected as evaluation indicators in this paper, and the comparison algorithm did not explain the computational complexity and time complexity, so no supplementary experiments were conducted for this point.   However, it can be seen from the epoch that the convergence speed of the algorithm in this paper is better than that of the comparison algorithm.  
Suggestions 7:
The resolution of all images in the paper has been improved, the axes have been explained, and modifications have been noted in the paper.

We have tried our best to optimize and modify the article. Finally, thank you again for your suggestions, which are very helpful to the improvement of the overall quality of the paper.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Overall nice article no corrections needed. 

Author Response

Dear reviewer
Thank you for your approval of our article. This provides great encouragement for our scientific research. We look forward to academic exchanges with you in the future.
Wish you every success in the future.

Reviewer 3 Report

Comments and Suggestions for Authors

1. How does the experimental evaluation consider different scenarios or settings, such as varying data sizes, different user-item interaction patterns, or cold-start situations?

2. How does the algorithm perform in these different scenarios?

3. I have seen ,there is no comparison to the state-of-the-art of recommendation methods in terms of recommendation accuracy, coverage, scalability, and robustness.

4. How did the author prove the statistically significant improvements?

5. What evaluation metrics were used to assess the performance of the recommendation algorithm? This information is not presented in this paper.

6. How do these metrics capture the effectiveness and efficiency of the algorithm in providing accurate and relevant recommendations?

7. The author needs to emphasize the dataset and their features but none of the sections are found.

8. How were these datasets selected, and what are their characteristics in terms of size, sparsity, and diversity?

 

Comments on the Quality of English Language

Ok 

Author Response

Dear reviewer
Thank you very much for your comments and suggestions.
Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches.    We have studied comments carefully and have made correction which we hope meet with approval.    The modifications are explained in the paper in the form of notes.
The following is a reply to your suggestion:
Suggestions 1:
Your suggestion is of great help to us, so we added MovieLens-1M data set, which has a larger data set than MovieLens-100K, and carried out supplementary experiments, which were marked in section 4 of the paper. For users, the project interaction pattern is determined as the scoring pattern based on the data set. As for the cold start problem, it is not the main problem solved in this paper, so it is not supplemented.
Suggestions 2:
According to the performance of the algorithm in different data sets, supplementary experiments are carried out in the experimental part of the paper and marked.
Suggestions 3:
In terms of recommendation accuracy, RMSE and MAE are selected to compare with the comparison algorithm, and supplemented in the experimental part.
Suggestions 4:
Additional information on the statistical improvement is provided in tables 2 and 3 and in figures 16,17,18,19.
Suggestions 5:
The evaluation indicators are supplemented in section 4.2 of the paper.
Suggestions 6:
The evaluation indicators in the accuracy of the recommendation algorithm are supplemented in Section 4.2. Lower RMSE and MAE represent better recommendation effect and accuracy
Suggestions 7:
In Section 4.1 of this paper, the characteristics of the two datasets are supplemented separately
Suggestions 8:
The size and sparsity of the dataset are supplemented in Table 1.
We have tried our best to optimize and modify the article.    Finally, thank you again for your suggestions, which are very helpful to the improvement of the overall quality of the paper.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thanks for addressing the queries, your efforts are appreciated.

Reviewer 3 Report

Comments and Suggestions for Authors

I have gone through details responses but still motivation is very weak in introduction section 

There is a chance to enhance the quality in literature review section  

 

Comments on the Quality of English Language

ok