Analysis and Prediction of User Sentiment on COVID-19 Pandemic Using Tweets

Round 1

Reviewer 1 Report

As this paper covers the Tweet users’ sentiment on COVID-19, its topic is of high importance and the work is of some interest. The primary contributions could be summarized as the applications of different ML methods and conventional DL methods to predict the sentiment. This work aims at an interesting topic but its contributions to the society is limited to the utilization of exiting techniques. Therefore, I highly recommend the authors made more contributions to the development of DL structures or ML methods. Or at least, make clear the contributions to modifications of these methods towards your goal. Otherwise, the technique contributions are questionable and limited, since authors only adopt common and existing methods in their research.

Author Response

Author's Reply to the Review Report (Reviewer 1)

As this paper covers the Tweet users’ sentiment on COVID-19, its topic is of high importance and the work is of some interest. The primary contributions could be summarized as the applications of different ML methods and conventional DL methods to predict the users’ sentiment using Tweet data on COVID-19. This work aims at an interesting topic but its contributions to the society is limited to the utilization of exiting techniques. Therefore, I highly recommend the authors made more contributions to the development of DL structures or ML methods. Or at least, make clear the contributions to modifications of these methods towards your goal. Otherwise, the technique contributions are questionable and limited, since authors only adopt common and existing methods in their research.

 

Response to Reviewer 1

Thank you so much for your comments.  For your kind information, we have explored the existing DL and ML methods for analysing the users’ sentiment on COVID-19. For this, we have used two tweet datasets of the users and considered various preprocessing methods to prepare the dataset in order to use in DL and ML. More particularly, we have considered several feature selection techniques in order to select most important features so that DL and ML approaches can provide higher accuracy. As such, in a nutshell, we can say that we make a performance comparison among the ML and DL approaches by considering various feature extraction methods in order to analysing users sentiment using COVID-19 related tweets.   

Reviewer 2 Report

This research aims to evaluate user sentiment by creating ML and DL models that can effectively forecast sentiment and compare COVID-19 infection cases and COVID-19 associated tweets. From 1st April to 15th April 2020, we gathered data from Twitter using the search keywords CORONAVIRUS and COVID-19 from nine states of the USA. Some typographical errors should be removed.

A couple of references from this journal should be added in reference section. The following references should be added in reference section:

1.  M.K. Sharma, N. Dhiman, Vandana, V.N. Mishra, Mediative fuzzy logic mathematical model: A contradictory management prediction in COVID-19 pandemic, Applied Soft Computing, Vol. 105, (2021), 107285. DOI: https://doi.org/10.1016/j.asoc.2021.107285.

2. M.K. Sharma, N. Dhiman, V.N. Mishra, L.N. Mishra, A. Dhaka, D. Koundal, Post-symptomatic detection of COVID-2019 grade based mediative fuzzy projection, Computers and Electrical Engineering, Vol. 101, (2022), Manuscript id: 108028. DOI: https://doi.org/10.1016/j.compeleceng.2022.108028

Author Response

Author's Reply to the Review Report (Reviewer 2)

This research aims to evaluate user sentiment by creating ML and DL models that can effectively forecast sentiment and compare COVID-19 infection cases and COVID-19 associated tweets. From 1st April to 15th April 2020, we gathered data from Twitter using the search keywords CORONAVIRUS and COVID-19 from nine states of the USA. Some typographical errors should be removed.

A couple of references from this journal should be added in reference section. The following references should be added in reference section:

1.  M.K. Sharma, N. Dhiman, Vandana, V.N. Mishra, Mediative fuzzy logic mathematical model: A contradictory management prediction in COVID-19 pandemic, Applied Soft Computing, Vol. 105, (2021), 107285. DOI: https://doi.org/10.1016/j.asoc.2021.107285.

2. M.K. Sharma, N. Dhiman, V.N. Mishra, L.N. Mishra, A. Dhaka, D. Koundal, Post-symptomatic detection of COVID-2019 grade based mediative fuzzy projection, Computers and Electrical Engineering, Vol. 101, (2022), Manuscript id: 108028. DOI: https://doi.org/10.1016/j.compeleceng.2022.108028

 

Response to the comments of reviewer

Thanks for your valuable comments. We have added these reference papers.

Reviewer 3 Report

Comments on Article 1721392: Analysis and Prediction of User Sentiment on COVID-19 Pandemic using Tweets

 

Paper basically sound but needs some editing to be suitable for publication. It is generally well written and reads like a ‘how to do sentiment analysis’ paper. Using both DL and ML approaches is interesting and makes for a good comparison of these techniques.

 

NOTES

• Needs careful proofreading and rewrite in parts – a few typos and grammatical errors
• Consistency needed with terms like ‘dataset’
• Tables and figs need to be in the sections where they are discussed.

 

DETAILED COMMENTS ON PAPER

Abstract

• Provides a reasonable overview of the paper

 

1. Introduction

 

2. Related Works

 

Seems thorough although there are other papers that could be referenced (see below)

 

3. Methodology

 

3.3 Sentiment Analysis

Use of textblob library explained. Are all sentiment values in the paper based on this? How accurate is textblob? Textblob has other functions that provide positive and negative polarity

 

opinion = TextBlob(tweet.text, analyzer=NaiveBayesAnalyzer())

opinion.sentiment.p_pos and opinion.sentiment.p_neg can be obtained using Naïve Bayes

 

 

4. Experimental Validation

 

Results generally support research findings

 

5. Discussion and Conclusion

 

Anticipation of working on other SM types is interesting. Reddit might be a better candidate since data is mainly text based like Twitter and thus more suitable for NLP

 

Additional References?

Paper does have a considerable literature review already with 52 cited. However, a quick search from google scholar listed these refs that might be relevant. There are obviously many more papers in this rich area of research.

• Neogi, A. S., Garg, K. A., Mishra, R. K., & Dwivedi, Y. K. (2021). Sentiment analysis and classification of Indian farmers’ protest using twitter data. International Journal of Information Management Data Insights, 1(2), 100019.
• Shofiya, C., & Abidi, S. (2021). Sentiment Analysis on COVID-19-Related Social Distancing in Canada Using Twitter Data. International Journal of Environmental Research and Public Health, 18(11), 5993. Shofiya, C., & Abidi, S. (2021). Sentiment Analysis on COVID-19-Related Social Distancing in Canada Using Twitter Data. International Journal of Environmental Research and Public Health, 18(11), 5993.
• Naseem, U., Razzak, I., Khushi, M., Eklund, P. W., & Kim, J. (2021). COVIDSenti: A large-scale benchmark Twitter data set for COVID-19 sentiment analysis. IEEE Transactions on Computational Social Systems, 8(4), 1003-1015.

Comments for author File: Comments.pdf

Author Response

Author's Reply to the Review Report (Reviewer 3)

 

Paper basically sound but needs some editing to be suitable for publication. It is generally well written and reads like a ‘how to do sentiment analysis’ paper. Using both DL and ML approaches is interesting and makes for a good comparison of these techniques.

 

NOTES

1. Needs careful proofreading and rewrite in parts – a few typos and grammatical errors

Ans: Thanks for this comments. We have revised the paper to fix these kinds of errors.

2. Consistency needed with terms like ‘dataset’

Ans: We have fixed this.

Tables and figs need to be in the sections where they are discussed.

 Ans: Thanks for this suggestion. We have fixed this. Please refer to Tables and Figures.

DETAILED COMMENTS ON PAPER

 

1. Related Works

 

Seems thorough although there are other papers that could be referenced (see below)

Ans: Thank you so much for your valuable comments. We have added these papers.

 

1. Methodology

 

3.3 Sentiment Analysis

Use of textblob library explained. Are all sentiment values in the paper based on this? How accurate is textblob? Textblob has other functions that provide positive and negative polarity

opinion = TextBlob(tweet.text, analyzer=NaiveBayesAnalyzer())

opinion.sentiment.p_pos and opinion.sentiment.p_neg can be obtained using Naïve Bayes.

Ans: It depends. Textblob provides better accuracy with more formal language usage, while it provides poorer accuracy with emoji, slang, etc. Since in Twitter users typically tweet with text, as such we have used Textblob.

 

1. Discussion and Conclusion

Anticipation of working on other SM types is interesting. Reddit might be a better candidate since data is mainly text based like Twitter and thus more suitable for NLP

Ans: Thanks for this comment. We plan to use Reddit to extend this work in future.

Round 2

Reviewer 1 Report

As the paper is will revised， I suggestion acceptance.