Neural Network for Sky Darkness Level Prediction in Rural Areas
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
Thank you so much for addressing an interesting research topic. I highly recommend the paper a major revision, or for re-submission in other related journals; the paper contains little information adequate to justify publication in sustainability. The paper is a very descriptive, simple analysis and it simply does not apply existing knowledge in a new context - country-Spain.
-- Spain as a place limitation is mentioned only 2 times in the manuscript...more information needs to be added to this concern to enrich the values of the limitations.
- The keywords were not dealt with specifically in terms of coverage in references and literature related to them...they need careful review, especially some words were not mentioned e,g, skywatching and some were mentioned once or twice in the research e.g. depopulation and Sky quality meter ....etc.There is an urgent need for more linked references to embrace the research contexts, methodology, and findings related to the given keywords.
-- The introduction does not explain the gap in the literature, and the author(s) do not provide enough information to make a good theoretical background.
-- The paper's argument does not build on an appropriate base of theory, concepts, or other ideas.....the used Multi-Layer Perceptron (MLP) model is very interesting but needs more time editing and clarifications in addition to the used Levenberg-Marquardt algorithm.
-- The results, conclusions, and implications for research, practice, and/or society are presented clearly and analyzed appropriately if extensive corrections are made to the paper
-- Reviewing the weak conceptual framework established at the outset and running throughout all sections could significantly reconstruct and improve the manuscript while clarifying potential fit in the journal in the future.
-- -- The author (s) provided a lack of discussion. The contributions of this study are fairly superficial. The “so what” answers of this paper remain unclear and quite obscure.
Comments on the Quality of English Language
Moderate editing of English language required
Author Response
"Please see the attachment"
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors
This paper has developed a neural network using the Multilayer Perceptron (MLP) model for predicting the night sky brightness (NSB) in rural areas. The authors' writing is fluent, but the content of the paper is relatively simple and lacks a more in-depth analysis. The main contribution of the authors is more like to present another case study in comparison to the Tokyo case, with no improvements made to the application of the MLP model itself. Specific suggestions are as follows:
1. It is recommended that the authors make certain improvements to the MLP model in response to specific application scenarios and compare the results of these improvements in the paper to enhance its innovativeness. Additionally, the authors are advised to include a model comparison, assessing the relative performance of the current model against other predictive models, such as other neural network models, regression models, decision trees, and random forests. This can be accomplished by comparing predictive errors and accuracy.
2. Figure 3 may not be necessary, as in practical applications, interfaces are typically implemented through web pages or mobile apps rather than using a Matlab interface for user operation. It is suggested that the authors include images of the field setup where the data was collected, as well as images and performance parameters of the sensors used. This information would be more valuable for readers.
3. In Figure 4, there appears to be a significant discrepancy between the predicted and actual values, and many predicted values are below the x-axis and not displayed. Can more data be shown?
4. The analysis in this paper is not very in-depth and does not further investigate the impact of weather conditions, moonlight, or the level of light pollution. For example, the authors could obtain weather data (such as cloud cover, humidity, wind speed, etc.) from local meteorological departments as input variables and integrate them into the model to assess the impact of these factors on the darkness of the night sky. This would make the model more comprehensive and potentially improve the accuracy of the predictions.
5. The comparison with the Tokyo case is not based on the same dataset. Although several performance indicators are better than those in reference [36], this may be due to regional differences rather than differences in the methods proposed, such as the data from Tokyo being more regular. In the comparison with reference [36], the paper simply discusses the differences with the Tokyo case but does not provide a quantitative analysis of the related content. It is suggested that the authors add a comparison of the methods proposed in this paper and the methods in reference [36] when applied to both the dataset of this paper and the Tokyo dataset.
6. The authors are encouraged to include an ablation study to analyze the contribution of various components in the model to the final results.
7. It is recommended that the authors conduct a sensitivity analysis to determine which input variables have the greatest impact on the model's output. This will help understand the decision-making process of the model and may reveal key drivers in the dataset.
8. The authors trained the data so that users can obtain night sky brightness data by inputting dates and times. In fact, the patterns in the data also change year by year. Has the author considered the overall changes between different years in the analysis?
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear Authors,
The manuscript entitled "Neural network for sky darkness level prediction in rural areas" is very interesting, and the issues it deals with are current. However, it still has some issues that need to be addressed. Below is the list of suggestions for manuscript enhancement:
The introduction offers a detailed historical framework for rural depopulation and industrial migration. While important, this historical perspective may overshadow the study's immediate goal of projecting night sky brightness for astronomical tourism.
Despite briefly mentioning previous models of night sky brightness, the introduction fails to critically assess their limitations or their potential unsuitability for rural areas. This oversight leaves a gap in understanding the need for a new model.
The introduction references the use of neural networks in predictive modeling but lacks an explanation of why neural networks are particularly well-suited for predicting night sky brightness or how they compare to other modeling techniques.
Certain points, like the influence of various factors on night sky brightness and the necessity of a neural network, are repeated or closely related. Streamlining these discussions would make the introduction more concise and focused.
Although the introduction mentions potential benefits for astro-tourists and astronomers, it does not discuss how these benefits will be practically realized or their potential impact on rural communities and tourism infrastructure.
The methodology relies on the Sky Quality Meter (SQM) and SG-WAS photometers. However, it does not address potential calibration issues, device accuracy, or how differences between devices were managed.
The results section highlights a strong correlation between actual and predicted values; however, it does not sufficiently address potential limitations of the neural network model. Specifically, it lacks discussion on the model’s sensitivity to variations in input data and the potential variability in prediction accuracy under different conditions.
The discussion compares the developed model with only one other model (C-Sánchez et al.), which may not provide a complete evaluation of its performance. A broader comparison with additional models would offer a more comprehensive assessment of the new model's relative advantages and disadvantages.
More detail is needed for the proposed future research direction of creating meteorological stations and adding environmental factors into the neural network. The recommended approaches for integrating these variables and their expected impact on model performance are not adequately articulated, leaving the reader with questions concerning practical implementation.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe study developed neural network using the Multilayer Perceptron (MLP) model to predict the darkness value of the night sky in rural areas,which has certain research significance.but there are the following problems.
1. Line 223: What is the calculation formula for R2 and MAE?
2. Section of Data collection:Please describe the data features of the training and validation sets using a table? For example, maximum value, minimum value, average value, etc.
3.Line 172: Why was the recorded data set divided into training (75%) and validation (25%) data?
4. Section of Results and discussion:Suggest the author list how the neural network is transmitted from input to output? Please list the transfer functions.
Comments on the Quality of English LanguageI considered that English language should to be modified appropriately.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThank you for your efforts. The manuscript has been updated and improved. Only one comment to.be considered.
1. Clarifying the manuscript contribution to both academic, decision makers, local communities, industry, theory and practices are required.
Comments on the Quality of English LanguageMinor English proofread needed for the new version.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors have addressed the reviewers' concerns, making proper revisions to the manuscript, and committed to future research for model refinement. The paper now meets the publication criteria.
Author Response
Thank you for your help in the revision process. The final version of the manuscript has been uploaded to the platform. Best regards.