Towards Hyper-Relevance in Marketing: Development of a Hybrid Cold-Start Recommender System

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Interesting topic and results, but poor work with references, citations are only in Chapter related works, citation are missing in rest of the article

 

abstract 

 - its too long, you should make shorted description of 4 used systems

 - add some numerical values (result, size of datasets)

 

References

 - only 16, ad at least 24 more (40 in total), at least 10 from them years 2021-23

 - generally you use citation only in Related works chapter, but you must use citation in whole rest of article, when you use some knowledge, which is not your own idea

 

Introdution

 - you have only one citations in this chapter, you must cite every fact what you are saying in this chapter (about 10 citations in this chapter)

 

Related works

 - this chapter is OK, maybe add a few related works from 2023

 

chapter 3.1

 - missing citation about sources of dataset

 

Results

 - very interesting, good work with figures and tables

 

Discussion

 - this chapter is compleatly missing

 - you must compare your results with results of related works

 

Conclussion

 - its OK

 

Figure 2 

 - can be removed, not interesting

Author Response

Dear Reviewer,

Thanks for your attention to detail and meaningful comments.

We've done the following changes:

• "citations are only in Chapter related works, citation are missing in rest of the article"

  • We've added citations in the introduction and in the rest of the document

•  

  abstract - its too long, you should make shorted description of 4 used systems - add some numerical values (result, size of datasets)

  •  

    We've reduced the abstract and included numerical values as datasets details and obtained results

• References  - only 16, ad at least 24 more (40 in total), at least 10 from them years 2021-23

  • We've added several references, having a total of 41 now, most of them from the last 3 years

 

• generally you use citation only in Related works chapter, but you must use citation in whole rest of article, when you use some knowledge, which is not your own idea
  • We've added citations in the introduction and in the rest of the document
• Introdution - you have only one citations in this chapter, you must cite every fact what you are saying in this chapter (about 10 citations in this chapter)
  • We now have 16 different citations in the introduction
• Related works - this chapter is OK, maybe add a few related works from 2023
  • We've added 3 works from 2023 to improve the study as suggested
• chapter 3.1 - missing citation about sources of dataset
  • The dataset is owned by a private company, it is not publicly available
• you must compare your results with results of related works
  • We've compared with previous work, that was like a baseline for us. As it is a hybrid approach, and part of a private organization, the only possibility of comparison was with the baseline, to ensure a fair comparison. 
• Figure 2 - can be removed, not interesting
  • Removed

Thank you one more time,

Best regards,

The authors

Reviewer 2 Report

Comments and Suggestions for Authors

This study is interesting and of great practical value, the theoretical analysis is relevant and comprehensive, the methodology is suitable with the question, the results are clearly reported, and the conclusions are well supported and the implications are reasonable. There are still several minor issues to be addressed.

1. There should be adequate references in the introduction to support your arguments.

2. There are several kinds of RS systems introduced in "2. Related Work", each of them are clear, but it would help if there is a table or graph to illuminate the similarities and differences among them and the evolution logic of their development.

3. Some more specific suggestions for enterprises or managers will be great.

Author Response

Dear Reviewer,

Thanks for your attention to detail and meaningful comments.

We've done the following changes:

• 1. There should be adequate references in the introduction to support your arguments.
  • We've added 16 different citations in the introduction

• 2. There are several kinds of RS systems introduced in "2. Related Work", each of them are clear, but it would help if there is a table or graph to illuminate the similarities and differences among them and the evolution logic of their development.

  • We've added a comparative table that summarizes the related work

•  

  3. Some more specific suggestions for enterprises or managers will be great.

  • We've added to the conclusions

Thank you one more time,
Best regards,
The authors

Reviewer 3 Report

Comments and Suggestions for Authors

This study proposes a hybrid recommending system for cold-start consumers by integrating four individual systems. It adopted a data set from a Portuguese company, E-goi, which serves as a case study of the proposed approach. I have one primary concern for this work and several minor suggestions. 

 

My primary concern is the weightings of the four subsystems are assumed to be  15 % to the 1CR, 15 % to the 2CR, 45 % to the CR, and 25% to the MR. Why and how to generate this setting is unavailable. Besides, the authors claimed the robustness of this work; it seems beneficial to apply different settings to evaluate its robustness.

Several minor suggestions are as follows:

(1) The abstract is lengthy;

(2) The assumptions and limitations of the hybrid model should be discussed in the Conclusion;

(3) Certain terms require further explanation, such as TF-IDF and one hot encoding;

(4) On page 7, lines 269 & 273 need to specify the figure.

Comments on the Quality of English Language

The quality of the English language requires moderate editing.

Author Response

Dear Reviewer,

Thanks for your attention to detail and meaningful comments.

We've done the following changes:

• "My primary concern is the weightings of the four subsystems are assumed to be  15 % to the 1CR, 15 % to the 2CR, 45 % to the CR, and 25% to the MR. Why and how to generate this setting is unavailable."
  • We've added the following explanation: "Our study concluded that the 1CR and 2CR are associated with a higher performance when its weight is lower than the remaining systems. The CR performs better at a higher weight and the MR performs well at a medium weight."

• Besides, the authors claimed the robustness of this work; it seems beneficial to apply different settings to evaluate its robustness.
  • We appreciate your comment, we're still improving the recommender system, we'll apply different settings to evaluate the robustness in future work

• (1) The abstract is lengthy;
  • We've reduced it
• (2) The assumptions and limitations of the hybrid model should be discussed in the Conclusion;
  • Added to the conclusion
• (3) Certain terms require further explanation, such as TF-IDF and one hot encoding;
  • Added to the respective sections
• (4) On page 7, lines 269 & 273 need to specify the figure.
  • Thank you, it was a mistake, we've corrected it

Thank you one more time,
Best regards,
The authors

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

All my recommandations were solved, I am OK with the current version of the article

Author Response

Thank you very much

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed the mentioned issues. I suggest to accept it.

Comments on the Quality of English Language

Moderate editing of English language is required.

Author Response

Thank you very much