An Effective ELECTRA-Based Pipeline for Sentiment Analysis of Tourist Attraction Reviews
Round 1
Reviewer 1 Report
The article is extremely interesting. However, I was slightly preoccupied by the plethora of authors (no less than 4) for 12-page article.
Author Response
We have rechecked the references carefully.
Reviewer 2 Report
1. In line 13, perhaps the authors can add more information such as “in order to solve … issue, this paper proposes what method….” to make the abstract clearer.
2. All acronyms should mention their full names when firstly appear in the text.
3. The title of Section 3 should be changed to Methodology.
4. The manuscript should be proof read, otherwise it is difficult to understand and read. Please re-check the grammar and spelling.
Author Response
1.In line 13, perhaps the authors can add more information such as “in order to solve … issue, this paper proposes what method….” to make the abstract clearer.
Response: In line 13, information has been added to the paper to make the abstract more clear.
2.All acronyms should mention their full names when firstly appear in the text.
Response: All acronyms were carefully checked and the full names were added one by one.
3. The title of Section 3 should be changed to Methodology.
Response: This has been modified.
4. The manuscript should be proof read, otherwise it is difficult to understand and read. Please re-check the grammar and spelling.
Response: Grammar and spelling have been re-proofed and checked to make the paper easier to understand and read.
Reviewer 3 Report
Dear authors,
The paper would be very interesting to the targeted readership. I would recommend the pulication after moderate English changes.
Kind regards,
Xiu
Author Response
Grammar and spelling have been re-proofed and checked to make the paper easier to understand and read.
Reviewer 4 Report
The work of "An Effective ELECTRA-Based Pipeline for Sentiment Analysis 2 of Tourist Attraction Review" by Hui Fang, Ge Xu , Yunfei Long and Weimian Tang lies in the field of Computing and Artificial Intelligence and probably focuses on developing of shallow machine learning for sentiment analysis of Big (text) data. The proposed a new ELECTRA-based pipeline model that demonstrates the high effectiveness on sentiment analysis, for example, a tourist attraction review. The research improve the methodological relevance and applicability of current sentiment analysis methods.
My main concernings are about why do the authors everytime tell us about analysis of tourist attraction review? What is special about this data? Is it just a big text data or not? Not going to offend somebody, but now it looks like the article was generated by AI or neural networks, in part.For example, while talking about "stopwords" authors just overwhelmed the text by countless repeat of the same phrases and senseless abbreviations. From the first reading I could not also understand any significance or originality of this work. And authors did not help me at all.
Usually, scientists use the inductive method in research. They go from the particular to the general. Therefore, any scientific work should strive to obtain general reliable patterns, laws, and not some arbitrary frequent cases.
Finally I think the abstract, introduction part and conclusions of the manuscript should be corrected and rewritten keeping in mind aforementioned comments.
Author Response
The characteristics of tourist attraction review data and the specificity of the classification task have been further explained, which in turn highlights the originality and significance of this work.The abstract, introduction and conclusion sections of the manuscript have been revised taking into account the above comments.
