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.
The author has incorporated all the suggestions. The manuscript is ready for publication.
[# PeerJ Staff Note - this decision was reviewed and approved by Jyotismita Chaki a PeerJ Section Editor covering this Section #]
Thanks for your revised submission. I recommend a few more revisions:
- The author claims that "....techniques are not suitable for handling nonlinear data" (lines 181-182); however, there are studies discussing the topic
-A few figures (for example, Figure 5) should be clearer (expanded).
- There are some inconsistencies in the figures' titles. For example, the first letters in Figure 6 are capitalized, which is not the case for other figures.
I'll recommend you make a few minor corrections recommended by the reviewers.
Use RES in lines: 46, 55, 60, 81, 83.
Add spaces before citation in lines: 93, 188, and 201 (before PCA).
Add space in Table 2 (in Feature column words)
Centralize all Equations
Use all lowercase letters other than the 1st letter in lines: 103, 109, 249, 274, 280, 283, 287, Algorithm 1 line 24,
Use all lowercase letters other than the 1st letter in Figure titles: Figure 3, 4, 5, 6(a), 6 (b), 6, 7, and 9
Use all lowercase letters other than the 1st letter in Table 3: Learning rate, Batch size, GRU hidden units, ResNeXt block configuration, Dropout rate, Weight decay.
no comment
no comment
no comment
This manuscript proposes a RXT-J hybrid forecasting model for enhanced solar power generation prediction. The authors have demonstrably improved the manuscript by incorporating an empirical model, addressing all reviewer queries, and providing convincing justifications for the revisions. Based on the strength of these improvements, I recommend accepting this manuscript for publication.
Please see the above comment
Please see the above comment
Please see the above comment
I agree with the assessment of two reviewers. A significant revision is needed before the manuscript can be considered for publication. Please review the comments attached and address them.
**PeerJ Staff Note:** Please ensure that all review and editorial comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**Language Note:** The review process has identified that the English language must be improved. PeerJ can provide language editing services - please contact us at [email protected] for pricing (be sure to provide your manuscript number and title). Alternatively, you should make your own arrangements to improve the language quality and provide details in your response letter. – PeerJ Staff
ResNext is already an acronym. So, how can the acronym “ResNext” have another acronym, “RXT”?
Replace ResNext with ResNeXt in all occurrences.
Write the abbreviation of any acronyms before using it, for example, JA in line 109.
Make all lower-case letters in the following lines:
Line 17: change Artificial Intelligence to artificial intelligence.
Line 18, 94, 328, 405: Change Gated Recurrent Unit to gated recurrent unit.
Lines 94, 328, 405: Use just GRU instead of Gated Recurrent Unit.
Line 20: Change Principal Component Analysis to principal component analysis.
Line 201: Use PCA instead of Principal Component Analysis.
Line 26: Change Normalized RMSE to normalized RMSE.
Line 26: Change Normalized MAPE to normalized MAPE.
Line 40: Change Renewable Energy Systems (RESs) to renewable energy systems (RESs).
Line 202: Change Exploratory Data Analysis (EDA) to exploratory data analysis (EDA).
Line 92: Change Hybrid Architecture for Projecting Renewable Energy to Hybrid architecture for projecting renewable energy.
Change Renewable Energy Systems (RESs) to RESs to Lines 46, 50, 54, 77, 81, 83, 403.
Authors must use an acronym for renewable energy forecasting in line 86 and use that acronym for all the occurrences in Lines 91, 271, 291, 295, (Algorithm 24), 299, 315, and 321.
Line 26: Change “Implementing PCA for Feature Extraction” to “Implementing PCA for feature extraction”.
Follow these to the rest of the paper, i.e., change to lowercase to all other parts of the paper.
Line 93, 94, and all other occurences: Cite ResNeXt and GRU.
Line 147: What is [25]?
The authors write the abbreviation in line 165. Write this in the first occurrence.
Rewrite the Tables 1, 2, 3, and 4 titles with all lowercase except the first word and acronym.
Rewrite all the Figure titles with all lowercase except the first word and acronym. Also, explain what HEM is in the Figure 1 title.
Cite Figure 3 to the following paper (original published work):
Almazroi, Abdulwahab Ali, and Nasir Ayub. "Online Payment Fraud Detection Model Using Machine Learning Techniques." IEEE Access 11 (2023): 137188-137203.
The authors must add citations for all existing methods in Table 4.
Overall, the structure of this paper is well organized. The preprocessing of this research is clear and relevant.
This paper made research using ResNext-based GRU with Jaya optimization to strengthen the stability of renewable energy systems (RESs) and transition to sustainable energy resources to address climate change and save fossil fuels. They claimed that this study combined green and nature-inspired approaches to develop resilient energy systems that can resist disasters and aid in disaster recovery efforts. Their conclusion addressed the research question. Overall, the structure of this paper is well organized. The preprocessing of this research is clear and relevant.
However, the simulation and result section is very explanatory. A more concise explanation could be helpful for the readers.
The authors have implemented a new model and compared its performance with some existing models. However, they should also analyze the complexity of their proposed architecture in terms of parameters and time required and compare it with existing methods. This will help to determine how complex their architecture is compared to existing methods. Also, the ResNeXt is an old model to use. Many renowned models have been published recently, like the Newer version of SqueezeNext, MobileNetV2, CMT, and many more. The authors could use these recent models instead of ResNeXt.
The implementation of ResNext architecture with GRU has already been introduced. Authors must explain their architectural novelty.
The authors published another research work with a similar architecture (ResNeXt-embedded Gated Recurrent Unit (GRU) model (RXT) with Jaya optimization) and procedure (PCA to extract features) just for a different application, which is cited as:
Almazroi, Abdulwahab Ali, and Nasir Ayub. "Online Payment Fraud Detection Model Using Machine Learning Techniques." IEEE Access 11 (2023): 137188-137203.
The authors must clearly explain the architectural difference between ResNeXt-GRU-J and RXT-J.
The authors must explain more about split path cardinality.
Explain the proposed architecture novelty clearly.
Pay more attention to using uppercase and lowercase letters all over the document.
This manuscript introduces an empirical model to improve the forecast of SOLAR power generation by proposing an RXT-J hybrid forecasting model.
The authors believe that (1) the RXT-J model is the most accurate for solar PV, with error rates as low as 6.51 percent for normalized RMSE and 4.34 percent for normalized MAPE; and (2) tests against more complicated models like ResNet and decision trees demonstrate the true precision of the hybrid approach.
The paper is well organized and written and covers recent publications. I would recommend this paper. The manuscript presents an interesting and positive contribution by using a statistical hybrid model in the field of forecasting solar power generation. However, there are several key aspects that should be addressed before considering publication.
1. The concept of the renewable energy (RE) spectrum is big, including solar, wind, tidal, geothermal, bio, etc. The current manuscript is mainly forecasted using solar and a bit of wind energy data (e.g., Table 2) and not in the other areas of RE. The authors acknowledged that the study is solar energy-focused in lines 324–325 and in the conclusion part (lines 416–417) and that their future research will accommodate some of the dynamics of future renewable energy sources. So the author should rename it suitably ('Solar Energy' might work, including the title) in the entire manuscript to avoid confusion and to the readers' benefit.
2. Some of the figures are in low resolution and need to be improved (e.g., Figure 3).
3. Table 5 needs attention to which row or column is a model or algorithm.
4. Figure 9 has a blank of data between 29/4/22 and 17/05/22. Is that a loss of data, or was it not recorded? How are those values incorporated into the model, and what's the impact of the 'loss of data' in forecasting using the hybrid model?
5. Figure 10 is confusing to read. Choosing a better color might help. Why is the prediction less than the actual high pick in the figure?
6. The acronym 'PV' (Line 28) is not defined anywhere in the entire manuscript. It is always expected to elaborate on the acronym before it is used anywhere.
7. To make a well-connected and coherent manuscript, the authors need to revise the abstract and conclusion after all the revisions are completed.
The methods described here have sufficient detail and information.
The research question is well defined, relevant, and meaningful. The authors did an excellent work in literature reviewing.
It is usually expected to maintain steps in the validation of the forecast: 80:10:10 (training: testing: validation). The authors used 80:20 (training: validation) in this study. Please explain why you used this method. How is it impacted in validation when you have a loss of data in the input?
Please see the above.
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.