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.
I confirm that the authors have addressed all of the reviewers' comments and the manuscript is ready for publication.
[# PeerJ Staff Note - this decision was reviewed and approved by Jyotismita Chaki, a PeerJ Computer Science Section Editor covering this Section #]
Clear and professional English used throughout.
Literature references, sufficient field background/context provided.
Original primary research within the Aims and Scope of the journal.
Conclusions are well stated, linked to original research questions & limited to supporting results.
I appreciate the authors' effort to address my comments.
The manuscript is clear and understandable. The authors introduce the topic very well.
The literature and the references provided are sufficent.
The line of research is clear and understandable. This line cover the aims and scope of the journal.
The results are demonstrated and justified well. I indicate an ablatio study to understand your work's weak/strong components.
The paper needs significant modifications.
[# 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 rebuttal 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 rebuttal 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 rebuttal letter. Directions on how to prepare a rebuttal letter can be found at: https://peerj.com/benefits/academic-rebuttal-letters/ #]
1. The manuscript is well-written, with no typos except for Tables 1 and 2. They do not appear. You may need to re-label the tables.
2. If the class label (Goal 1 or 2) is based on a one-dimensional scalar input value (i.e., CN or CO), why are ML models used? Isn't it better to use a threshold or compute the correlation to show the effectiveness of CN and CO scores?
3. What is the "Tree" classifier? Can you cite it?
4. What are the parameters of the ML models?
5. How are the ML models implemented?
6. Since training is applied during cross validation, I wonder why FinBERT is not finetuned using training data and Transfer learning techniques.
7. Fig.2 and 3, which show the relationship between NB accuracy and volume, could be combined in one stacked bar chart.
8. Could you please explain why VADER and FinBERT were chosen but not the Loughran-McDonald dictionary in the Sentiment Analysis Test Result Subsection?
9. Can you please explain why the opening hours division is better than the natural hours division?
10. Can you show how the experiments concluded that the opening hours division is better than the natural hours division?
11. The authors may have to show the performance of other related approaches on the same testing dataset.
12. Can you please show the performance of TN, TO, NN, and NO separately?
13. Can you explain why NB got the maximum accuracy?
14. Why RandomForest got the highest F1?
15. The authors may have to balance the four stock datasets using data augmentation or resampling, and show the results.
The article is fluent and readable.
The background section seems complete, although I think it is important to include alternative references such as (https://www.mdpi.com/2078-2489/13/11/524). These papers focus purely on numerical data, but it is important to cite them.
T_weighted, N_weighted, and the other notations, such as Goal_I, are very well explained.
The primary motivation that led the authors to propose the measures that are the heart of the paper lies in the fact that most previous research adds up the text sentiment score on each natural day and uses that aggregate score for that aggregate score to predict various headline trends.
The underlying design seems reasonable, and it all seems to work well.
The research questions seem to be well and accurately asked.
However, there has not been a rigorous investigation. First of all, the figures you use should have a better resolution.
Second, vader is a bad and old model, especially in recent stock and stock sentiment.
It would help if you did an ablation study by removing portions of your architecture.
The ablation study will be the key to your research question. Despite the various metrics you put out, there is something that needs to sound right scientifically. An ML model outperforms the performance of FinBERT. This fact seems very strange.
Indeed, the correlation between data volume and accuracy is correct, but I do not think it is some hot water discovery.
1. Clear and unambiguous, professional English used throughout.
2. Literature references, sufficient field background/context provided.
3. Professional article structure, figures, tables. Raw data shared.
1.Original primary research within Aims and Scope of the journal.
2.Methods described with sufficient detail & information to replicate.
1.All underlying data have been provided; they are robust, statistically sound, & controlled.
2.Conclusions are well stated, linked to original research question & limited to supporting results.
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.