Deep Reinforcement Factorization Machines: A Deep Reinforcement Learning Model with Random Exploration Strategy and High Deployment Efficiency

Round 1

Reviewer 1 Report

In this paper, a deep reinforcement factorization machine (DRFM), which is the successful result of applying reinforcement learning to recommendation system. From my view, this paper is well organized and the proposed method is valuable for this research filed. After reviewed this paper, there are some questions and suggestions as follows.

1. Abstract section needs to be re-drafted to be self-contained means it has to clearly show the hypothesis, methodology, techniques and tools used, and the results obtained.
2. What assumptions authors made during the simulation phase of this research work? If there is any.
3. All figures need to be enhanced in terms of quality and resolution.
4. The literature review is poor in this paper. You must review all significant similar works that have been done. Also, review some of the good recent works that have been done in this area and are more similar to your paper.
5. It is necessary to talk about the role of the parameters of the proposed algorithm in a separate section. For example, which parameters are responsible for controlling exploration and exploitation?
6. It is necessary to experimentally analyze the proposed algorithm in terms of time consumed and compare with other algorithms.
7. What are the advantages and disadvantages of this study compared to the existing studies in this area?
8. There are many grammatical mistakes and typo errors. 
9. Write a pseudocode in standard format for the proposed algorithm.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments for author File: Comments.docx

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

1. In the literature, it will be more appropriate to compare the proposed method with their previous research (GAFM model) in order to solidify the contribution.
2. The performance of DRFM was better than GAFM model. But the run time seems not good. The learning time should be considered.
3. Recently, there were some new deep Factorization Machines. The authors should do the literature review and make comparison with these new models.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

This paper proposed a Deep Reinforcement Factorization Machines (DRFM) which is a fully upgraded deep reinforcement learning recommendation model. The proposed model combined the authors’ previous proposed model (Gate Attentional Factorization Machines (GAFM)) and reinforcement learning for improving accuracy and running speed. The authors compared their proposed model with several traditional recommendation system models on a variety of data sets. The results show that the proposed model outperforms the others.

Overall, the authors’ idea is interesting. However, the performance of the proposed was improved a bit compared with their previously proposed model (GAFM) in terms of running time and accuracy. It would be better if the authors could give more results that can show the explicit difference in performance between the current proposed model and the previously proposed model. Moreover, the number of epochs is quite low. Are 15 epochs the best? What would happen if the number of epochs is increased, such as 50 or 100 epochs?

In addition, the authors must be rewritten Section 3-A by using new descriptions and sentences that are different from the description and sentences used in their previous work [1].

[1] Yu, Huaidong, Jian Yin, and Yan Li. "Gate Attentional Factorization Machines: An Efficient Neural Network Considering Both Accuracy and Speed." Applied Sciences 11, no. 20 (2021): 9546.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Good revisions have been made in the paper and the revised version has the necessary qualities for acceptance compared to the previous version. In my opinion, the article is acceptable in its current form.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors had follow the requirements. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

The authors addressed all my concerns. However, before it possibly is accepted, the authors should correct some little mistakes, such as the definition of variables of equations (2) and (3), which seems the subscript of "g" in the description is missing. There might be other little mistakes, please recheck the whole paper carefully.

Author Response

Please see the attachment.

Author Response File: Author Response.docx