All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Kindly check the manuscript for minor language edits.
[# PeerJ Staff Note: Although the Academic Editor is happy to accept your article as being scientifically sound, a final check of the manuscript shows that it would benefit from further English editing. Therefore, if you can identify further edits, please work with our production group to address them while in proof stage #]
Kindly improve the English language writing and correct the typographical mistakes.
[# PeerJ Staff Note: The Academic Editor has identified that the English language must be improved. PeerJ can provide language editing services - please contact us at copyediting@peerj.com for pricing (be sure to provide your manuscript number and title) #]
no comment
no comment
no comment
The following questions should be addressed by the authors:
1) line 280 typo error
2) line 301 typo error
3) line 320, was there any statistical analysis carried out?
4) line 363, how many languages and list them?
5) line 405, what are the input features for the various ML classifiers? covered in section 5.2.2 but not in section 4!
6) line 408 and 409, svm is a binary classifier and not a multiple class classifier.
7) line 489, which type of word embedding is used, is it CBOW or count based?
8) line 529, how is the polarity score computed?
1. The concept is clear.
2. In line no. 280, "sophisti. cated" - spelling mistake.
3. In figure 2, both Phase 2 and 3 consist of lexicon building block (you may indicate flow of phase 2 and 3 work in this diagram)
4. In line 383, 4.4.2 Phase 2: Lexicon generation from labeled data - you mentioned like that. But in figure 2, phase 2 input is taken from both labeled and unlabeled like that. Rectify this.
1. To improve the readability, the author requires to provide a good data flow/ sequence diagram in understanding the “Service Level Agreement”
1. In the introduction, the findings of the present research work should be compared with the recent work of the same field towards claiming the contribution made. , kindly provide several references to substantiate the claim made in the abstract (that is, provide references to other groups who do or have done research in this area).
2. Try to concise the conclusion.
3. Discuss the future plans with respect to the research state of progress and its limitations.
1. In the Introduction section, the drawbacks of each conventional technique should be described clearly.
2. You should emphasize the difference between other methods to clarify the position of this work further.
3. The Wide ranges of applications need to be addressed in the Introduction
4. Add the advantages of the proposed system in one quoted line for justifying the proposed approach in the Introduction section.
Kindly modify the paper as per the reviewers comments and suggestions.
[# PeerJ Staff Note: It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors are in agreement that they are relevant and useful #]
[# PeerJ Staff Note: Please ensure that all review comments are addressed in a response letter and any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate. It is a common mistake to address reviewer questions in the response letter but not in the revised manuscript. If a reviewer raised a question then your readers will probably have the same question so you should ensure that the manuscript can stand alone without the response letter. Directions on how to prepare a response letter can be found at: https://peerj.com/benefits/academic-rebuttal-letters/ #]
no comment
The discussion of the experiment needs to be more in-depth, and the content of the experiment needs to be more detailed and complete. Please analyze the advantages of the proposed method and supplement the related experiments.
no comment
1. Your most important issue
Q1: Are the methods presented in this paper the most advanced and original, what are the main challenges of this work, and these need to be further discussed in the relevant work.
Q2: The discussion of the experiment needs to be more in-depth, and the content of the experiment needs to be more detailed and complete. Please analyze the advantages of the proposed method and supplement the related experiments.
2. The next most important item
Q3: 3.3.2 What is the meaning of the PMI formula and its variables? Please give specific explanation.
Q4: 3.3.3 “After applying approach 2, our dataset contains around 38000 pseudo-labeled reviews. We then employ PMI and POS tagger in a similar way to phase 2. However, since this phase utilizes pseudo-labeled data instead of the true-label data, we set a higher threshold of 0.7 for the class label assignment.” Why is the threshold set to 0.7?
Q5: 5.1.1 “If the total polarity score of a review is below 0, we consider it as a positive prediction; if the final score is below 0, we consider it as negative;” In which case does the score below 0?
3. The least important points
Q6: Table reference error in section 3.1.2, Table3 should be changed to Table1.
Q7: 3.1.2 Incorrect data in Table1, the data in the last column of the last row in Table 1 should be 10410.
Q8: 5.1.2 “Among the three translated lexicons, VADER classifies 46.60% reviews correctly, while AFINN and Opinion Lexicon provide 48.65% and 41.95% accuracy, respectively.” The accuracy of AFINN in Table 4 is 31.65%, which does not agree with the description in the text.
Q9: CONCLUSION “We made both BengSentiLex and BengSwearLex publicly available for the researchers in Sen (2020)” This sentence lacks punctuation.
The article has been provided in clear and professional English, as well as a sufficient organized sections. The tables and Figures are also well created.
The approach seems to have a promising results. Although the approaches have been used in previous papers, they showed a good insight on the specific language (Bengali).
The authors have provided well-designed conclusion, as well as defining the data with good statistics.
Here is a few suggestion that can improve the article:
1- The scope of the "related works" section requires a more thorough details.
i) Some essential parts of the paper, such as "supervisory characteristics" (supervised methods, semi-supervised methods, and unsupervised methods) can be discussed.
ii) the following paper has also done a similar work. it is recommended to review and compare the method with the provided method in the paper. https://www.sciencedirect.com/science/article/abs/pii/S0950705120305529
iii) Section 3.2.1, the second paragraph is relatively discussed the related works, which can be replaced in the "related works" section.
2- For better referencing the readers, it is recommended to reference the Equations with numbers.
3- For better understanding of the readers, a translation for Figure 5 is required.
4- Also, the source of the dataset (Youtube reviews) is needed to be cited.
5- The compared methods are mostly traditional methods for the experiments. authors must provide state-of-the-art methods for comparisons. (having LSTM, Bi-LSTM, CNN, Attention model, BERT model, etc.) - though not all of them, but to compare some of them.
The general idea behind the paper is a novel approach for Bengali language and hope the authors find the comments useful and make the necessary changes for the acceptance.
No comment
No comment
No comment
line 231, Table 3 -> Table 1
line 235, numbers do not add up to the numbers tabulated in Table 1 -> last row, last column cell?
line 249, will translation sufficiently map all types of sentiments from one to another language?
line 254, reference for NMT
line 364, which is the machine learning model used in SGD algorithm? Similary in line 515.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.