Emoji, Text, and Sentiment Polarity Detection Using Natural Language Processing

Round 1

Reviewer 1 Report

The academic paper under review proposes an emoji-based framework for sentiment analysis that combines linguistic patterns of text and emojis to generate sentiment polarity. The authors note that virtual users generate a large volume of unbalanced sentiments on online crowd-sourcing platforms, which includes text, emojis, or a combination of both. The proposed framework aims to provide accurate sentiment analysis that can bring profits to various industries and services.

The study used 650 world-famous personages consisting of 168,548 tweets across the world to evaluate the proposed natural language processing framework. The authors found that the existence of emojis in sentiments can change the overall polarity of the sentiment. The CLDR name of emoji is utilized to evaluate the accurate polarity of emoji patterns, and a dictionary of sentiments is adopted for evaluating the polarity of text.

The authors evaluated the performance of three machine learning classifiers (SVM, DT, and NaïveBayes) for proposed distinctive linguistic features. The results showed that the proposed approach outperformed the other ML classifiers, particularly in SVM classifier.

Also, author need strengthen the problematisation part of introduction, clearly underline the area of their contribution after the literature review, and contrast the findings with the existing literature, and outline limitations and future research avenues.

Authors should consider more previous works (e.g., theoretical, conceptual, and empirical reviews) published in the literature. Authors should discuss the results and how they can be interpreted from the perspective of previously published studies:  DOI: 10.3390/joitmc8010012 , DOI: 10.5815/ijisa.2022.02.03 , DOI: 10.5815/ijitcs.2022.02.01 , DOI: 10.5815/ijmsc.2022.01.04

The study concludes that the proposed polarity detection generator achieved an exceptional perspective of sentiments present in the sentence by employing the flow of concept established based on linguistic features, polarity inversion, coordinated and discourse patterns, surpassing the performance of existing state-of-the-art approaches.

Paper is well-written and presents an interesting approach to sentiment analysis that combines text and emojis. The authors clearly stated the research question and objectives, and the methodology used is appropriate for the research question. The findings are presented clearly, and the authors have discussed the implications of the findings for sentiment analysis. The study's results provide valuable insights into sentiment analysis and its potential for various industries and services.

the study's limitations should be acknowledged, including the small sample size used for evaluation and the need for further research to validate the proposed framework's effectiveness. Nevertheless, the paper provides a promising approach for sentiment analysis that can benefit industries and services that rely on accurate sentiment analysis.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Reviewer 2 Report

Most researchers disregard emojis from the text while performing the sentiment analysis. However, incorporating emojis could present with a whole new dimension to sentiment analysis as textual sentiment analysis could represent something that is quite opposite to the presented emojis. In this regards, this paper demonstrated an interesting topic of classifying Emoji’s into sentiment analysis workflow. Overall the I am quite happy to see the methodology and the also the presented results.

However, the overall quality of the paper could be improved significantly with following:

1)      I can see that the impact of this research could be significant. However, at this stage, the author failed to clearly portray the research implication (i.e., how this research would benefit other researchers, industry etc.). It would be a good idea, to enlist the research impact within discussion area, by stating how the current body of knowledge is serving a particular area (e.g., movie review, product review, customer feedback etc.) and how this unique approach (i.e., integration of emojis to contextualize the overall sentiment analysis process) would help. This would serve as the rational for research (which is currently not clear).

2)      What are the limitations for this work? Please highlight the limitations within the conclusion area. Highlighting the limitations would allow propagation of future research in this area.

3)      This paper is ill-formatted and therefore, it was irritating towards the readers. For example, the header of Table 1 in page 3 and the rest of the table is in page 4. In another scenario, in one page you have the Table header of Table 3 (in line 263) and in another page you have the table??? Then in another page, you have huge empty spaces (e.g. line 366 to 376)

4)      Currently, there is serious flaw in how the paper is structured. In section 3 (proposed framework), the authors presented the system components (i.e., Tree generation, Parsing algorithm based on linguistic feature, pattern formation, emoji/text and final polarity evaluation). However, rather than placing these system component as 3.1, 3.2, 3.3, 3.4, the author used separate sections for each of these topics (i.e., section 4, 5, 6, 7). Because of this the sections appears to be completely disproportionate. For example, section 7 is just 10 lines???

5)      In Table 9, the second and third column is what? Is it classification error? Is it AUC? Is it Precision percentage/ Recall percentage? Is it Accuracy Percentage? What’s the evaluation metric? Please clarify. The author only mentioned result…what evaluation metric was used for this result?

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

1. Abstract should be stated clearer for highlighting the contributions.

2. Discussion section should based on the results to interpret and discuss. In addition, what are the advantages and contributions of the current research?

3. 6.2.1 sub-section seems unnecessary as there is no 6.2.2. It is suggested that the tile of 6.2 should be modified.

4. Please add more information to state the contributions of the proposed method in Discussion section.

5. Conclusions should be stated clearer for highlighting the contributions.

6. The manuscript should be proof read, otherwise it is difficult to understand and read. Please re-check the grammar and spelling.

7. Some literatures are obsolete, please update.

8. In terms of NLP, the following literature can be cited to improve the readability:

https://doi.org/10.1002/advs.202203990

https://doi.org/10.3390/app13031699

https://doi.org/10.1145/3561970

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Accept in present form

Reviewer 2 Report

All my comments and suggestions have been addressed. I have no further comments. Good Work!

Reviewer 3 Report

Accept in present form