Personalized Search Using User Preferences on Social Media

Round 1

Reviewer 1 Report

The authors propose a new search scheme to provide search results based on the user’s preferences 15 on social media. The idea seams interesting and novel, however, I have the following suggestions to improve its overall quality.

1.      The article should be thoroughly revised for clarity and grammatical accuracy. In the present form, they are numerous grammatical errors which hamper its readability.

2.      The abstract and conclusion should be improved to present the core contributions of the work and highlight the case study of “User Preferences” taken into account.

3.      TF-IDF is a general method, and we need to discuss why this method is adopted?

4.      How to prove the method of similarity determination, simple but effective?

5.      The authors mention that they have highlighted the existing approaches similar to their work. However, some of the latest literature is not included, please add, such as,

- A Topic Representation Model for Online Social Networks Based on Hybrid Human-Artificial Intelligence. IEEE Trans. Comput. Soc. Syst. 8(1): 191-200 (2021)

-A Public Psychological Pressure Index for Social Networks. IEEE Access 8: 23457-23469 (2020)

-SESM: Emotional Social Semantic and Time Series Analysis of Learners' Comments. SMC 2020: 4134-4139

-Zhihong Tian, Zhenji Zhang, Dongpo Xiao: Study on the Knowledge -Sharing Network of Innovation Teams using Social Network Analysis. ICEIS (2) 2011: 438-443.

 

Author Response

Dear Reviewer,

We would like to sincerely thank you for your attentive indications and good comments. Our manuscript was partially rewritten in order to reflect and complement your comments. Please refer to the attached file about the detailed revisions. 

Many thanks.

Jaesoo Yoo

Author Response File: Author Response.docx

Reviewer 2 Report

This paper investigates the personalized search using user preferences on social media and gives the comparison. It is well-written.

Author Response

Dear Reviewer,

We would like to sincerely thank you for your good comments. 

As you pointed out, we carefully reviewed the paper and corrected the typo and grammatical errors.

Many thanks.

Jaesoo Yoo

Reviewer 3 Report

1. The paper lacks of sufficient experiment with the state-fo-the-art works

2. There is a need for a discussion section to highlight the overall contribution of the paper

3. There is a need to improve the related works. For example check the following relevant papers:

Li, Jiuyong, et al. "Causal heterogeneity discovery by bottom-up pattern search for personalised decision making." Applied Intelligence (2022): 1-15.

Esposito, Marco, and Leonardo Picchiami. "A Comparative Study of AI Search Methods for Personalised Cancer Therapy Synthesis in COPASI." International Conference of the Italian Association for Artificial Intelligence. Springer, Cham, 2022.

Abu-Salih, Bilal, et al. "Toward a knowledge-based personalised recommender system for mobile app development." arXiv preprint arXiv:1909.03733 (2019).

- Also the paper might address the importance of semantic analysis and trust aspects when it comes to social media analytics, check the following works:

Abu-Salih, Bilal, Pornpit Wongthongtham, and Chan Yan Kit. "Twitter mining for ontology-based domain discovery incorporating machine learning." Journal of Knowledge Management (2018).

 

 

Author Response

Dear Reviewer,

We would like to sincerely thank you for your attentive indications and good comments. Our paper is partially rewritten in order to reflect and complement your comments. Please refer to the attached file about the detailed revisions.

Many thanks.

Jaesoo Yoo

Author Response File: Author Response.docx

Reviewer 4 Report

The proposed approach is interesting and worth consideration. Still, the paper has several issues regarding its consistency and completeness of the presented results, which should be taken into account before proceeding it to further stages.

-        Even if the paper presents a  description of the approaches in literature, it lacks a critical evaluation of the different  proposals that are not critically evaluated and compared with related work/state-of-the-art and are not identified and discussed the  drawbacks and limitations. Therefore, it is not easy to assess the real contribution of reviewed papers in the field and how the proposed solutions are efficient compared to related works. A clear assessment of the contribution of the different approaches should be given.

According to this perspective, further discussion has to be provided on  critical evaluation of contributions, the challenges produced, the possible benefits and future aspects of the AI techniques in the faced domain.

-       Another issue that should have been explored in the paper concern the adoption or proposal of ad-hoc topic detection methods to extract topics of discussion from social media and then catch contradictory comments.  In fact, personalized search using user preferences on social media could greatly benefit of the topics/events of discussion involving the users that could influence also users search. I think that the combination of topic detection and the solution proposed by the authors also exploiting Graph Convolutional network could enhance the effectiveness and accuracy of the approach. At least the authors should discuss Topic Detection on Social Media in the Related Work section. Also drawing some ideas as future work.  Accordingly, the Related Work Section should include also other references discussing how data from social media become essential to grasp people feeling, sentiment and opinion trough topic detection and modeling.  In particular, the authors should consider also  related approaches that consider the whole content of a tweet and not just the hashtag. See for example the following references:

o   https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IcgxFuIAAAAJ&cstart=20&pagesize=80&citation_for_view=IcgxFuIAAAAJ:35N4QoGY0k4C

 

o   https://ieeexplore.ieee.org/abstract/document/7814622

 

 

-        Authors needs to revise the experimental evaluation section and  improve the quality of the results representations, according to the following comments:
* A more extensive discussion of the experimental setting is required.

* The experimental data must be more extensively detailed, by also reporting insightful statistics.

 

 

Finally, there are minor spelling issues and typos. Please proofread and correct.

Author Response

Dear Reviewer,

We would like to sincerely thank you for your attentive indications and good comments. Our paper is partially rewritten in order to reflect and complement your comments. Please refer to the attached file about the detailed revisions.

Many thanks.

Jaesoo Yoo

Author Response File: Author Response.doc

Round 2

Reviewer 3 Report

Authors have addressed my concerns, I have no further ones. 

Reviewer 4 Report

The authors addressed all the issues and comments of my previous review. Therefore, the paper can be accepted as is.