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.
Dear authors,
The reviewers have commented on your paper. They indicated that it is now acceptable for publication in its present form.
[# PeerJ Staff Note - this decision was reviewed and approved by Claudio Ardagna, a 'PeerJ Computer Science' Section Editor covering this Section #]
Thank you for submitting the revised manuscript response to the feedback from the last review round.
I have carefully reviewed the response file containing your detailed answers to each point raised. I appreciate the comprehensive manner in which you addressed the reviewers' comments and revised the manuscript.
It can be accepted for publication.
Thank you for submitting the revised manuscript response to the feedback from the last review round.
I have carefully reviewed the response file containing your detailed answers to each point raised. I appreciate the comprehensive manner in which you addressed the reviewers' comments and revised the manuscript.
It can be accepted for publication.
Thank you for submitting the revised manuscript response to the feedback from the last review round.
I have carefully reviewed the response file containing your detailed answers to each point raised. I appreciate the comprehensive manner in which you addressed the reviewers' comments and revised the manuscript.
It can be accepted for publication.
Thank you for submitting the revised manuscript response to the feedback from the last review round.
I have carefully reviewed the response file containing your detailed answers to each point raised. I appreciate the comprehensive manner in which you addressed the reviewers' comments and revised the manuscript.
It can be accepted for publication.
It's good; no further comment.
It's good; no further comment.
It's good; no further comment.
Dear authors,
The reviewers have commented on your paper. They indicated that it is not yet acceptable for publication in its present form.
However, if you feel that you can suitably address the reviewers' comments, I invite you to revise and resubmit your manuscript.
Please carefully address the issues raised in the comments.
If you are submitting a revised manuscript, please also:
a) outline each change made (point by point) as raised in the reviewer comments
b) provide a suitable rebuttal to each reviewer comment not addressed
• The paper provided a clear overview of the dataset used in the study. It described the dataset's characteristics, including the number of records, columns, and the types of data included (such as fabric properties, manufacturing parameters, quality-related information, etc.).
• The paper adequately described the methods for data preprocessing and feature engineering. It outlined the steps to clean the data, handle missing values, and convert object-type properties into numerical representations. Additionally, it discussed the normalization of continuous data and the encoding techniques used for categorical variables like sin/cos encoding.
• The presentation of results in tables and figures was clear and easy to understand. Tables were used to present feature importance rankings for different AutoML tools, and figures were used to illustrate metric scores over time intervals, enhancing the readability and interpretation of the results.
• The chosen AutoML tools effectively addressed the research objectives, as they facilitated the training and evaluation of a regression task to predict fabric production quality. The paper discussed the performance of FLAML, AutoViML, EvalML, PyCaret, AutoGluon, H2O AutoML, and TPOT, providing insights into their capabilities regarding computational effort and predictive performance.
• The paper defined criteria for selecting the best-performing AutoML tools based on average predictive scores and average computational effort expended for each specific task and scenario. This approach was justified as it considered both predictive accuracy and computational efficiency, crucial factors in real-world applications.
• The paper adequately discussed the limitations of the experimental setup, such as the reliance on open-source AutoML tools and the need for domain expertise in collaboration with data scientists. It also mentioned potential biases, such as the lack of specific information about the dataset's distribution and the challenges of imbalanced datasets.
The paper demonstrated the robustness of the predictive models generated by AutoML tools in predicting fabric production quality. It evaluated the models' performance based on metrics like RMSE, MAE, R-squared, and MAPE, showcasing their effectiveness in capturing and predicting key quality parameters.
The comparisons between actual and predicted values were statistically sound and meaningful. The paper employed rigorous statistical analysis, including RMSE, MAE, R-squared, and MAPE, to assess the accuracy of the predictive models. This approach provided a comprehensive understanding of the model's performance in predicting fabric production quality.
The paper discussed the implications of the findings in the context of practical applications in the textile industry. It highlighted the potential for using machine learning, specifically AutoML, to automate and enhance quality control processes in textile manufacturing. The findings have practical implications for improving efficiency, reducing lead times, and increasing productivity in the industry.
• How did the paper address the practical implications of using AutoML for fabric production quality prediction, including potential cost savings, process optimization, and quality improvement?
• Discuss real-world deployment challenges or success stories related to implementing AutoML models in textile manufacturing environments.
• Did the paper adequately acknowledge and discuss the study's limitations, such as dataset size, data quality issues, or model interpretability challenges?
• Propose specific avenues for future research, such as exploring ensemble techniques combining AutoML models or addressing imbalanced data scenarios in fabric quality prediction.
• The paper provides a comprehensive overview of AutoML tools and their application in predicting fabric production quality, offering valuable insights into the advantages and limitations of different tools.
• Discussing feature importance and its impact on model interpretability adds depth to the analysis and understanding of the results.
• The paper effectively highlights the importance of collaboration between data scientists and domain experts in tackling real-world challenges in the textile industry.
To my opinion, the manuscript is suitable for publication after a minor revision towards this direction.
• Add keywords but avoid using the word the research paper's title.
• Draw picture one again or present it in parts because it’s difficult to read the content.
• Managerial implications are missing from the paper, add them.
• Add a result discussion and comparison section before the conclusion.
Remarks:
This paper is well-structured and well-written. It should be seriously considered. Overall, the paper is good for publication.
Add keywords but avoid using the word the research paper's title.
• Draw picture one again or present it in parts because it’s difficult to read the content.
• Managerial implications are missing from the paper, add them.
• Add a result discussion and comparison section before the conclusion.
Add keywords but avoid using the word the research paper's title.
• Draw picture one again or present it in parts because it’s difficult to read the content.
• Managerial implications are missing from the paper, add them.
• Add a result discussion and comparison section before the conclusion.
To my opinion, the manuscript is suitable for publication after a minor revision towards this direction.
• Add keywords but avoid using the word the research paper's title.
• Draw picture one again or present in parts because it’s difficult to read the content.
• Managerial implications are missing from the paper, add them.
• Add a result discussion and comparison section before the conclusion.
Remarks:
This paper is well-structured and well-written. It should be seriously considered. Overall, the paper is good for publication.
The manuscript is quite well-written, maintaining clear and professional language throughout. It conforms to the standard structure expected in the field of machine learning applications. The introduction provides a strong background, contextualizing the importance of integrating Industry 4.0 technologies like IoT with AutoML for fabric quality prediction. However, the use of acronyms without first defining them (e.g., TPOT, FLAML) can be confusing for readers unfamiliar with the specific terms. It is recommended that all acronyms be fully spelled out upon their first occurrence in the text. For example, clearly explain "TPOT" stands for "Tree-based Pipeline Optimization Tool" when it is first mentioned.
The research question is well-defined, relevant, and clearly fills an identified knowledge gap about automated solutions in predicting fabric quality. The manuscript details a rigorous evaluation of seven different AutoML tools (FLAML, AutoViML, EvalML, AutoGluon, H2O,PyCaret, and TPOT), providing a comprehensive comparison in terms of computational efficiency and prediction accuracy. However, the methods section could benefit from more detailed descriptions of the datasets used and the specific configurations of each AutoML tool to ensure replicability of the results.
The findings are robust, supported by statistical analyses and comparative metrics such as MAE, r-squared, etc. The conclusions drawn are appropriate and link well back to the research question, demonstrating the utility of EvalML and AutoGluon under different conditions. To strengthen this section, the authors might consider discussing the limitations of their study more explicitly, including any potential biases in the data and/or AutoML models.
- For the laborious hyperparameter tuning, there is a recent paper (see below) that automated the tuning process using Particle Swarm Optimization (PSO) algorithm, the authors may want to study it in their RELATED WORK section.
[Wang, Z., Li, J., Rangaiah, G. P., & Wu, Z. (2022). Machine learning aided multi-objective optimization and multi-criteria decision making: Framework and two applications in chemical engineering. Computers & Chemical Engineering, 165, 107945.]
- Overall, this is a good manuscript. I read it with great interest and I recommend it for "Accept with Minor Revisions". Addressing the comments will help clarify the methodology and findings, making the paper a more significant contribution to both the fields of textile manufacturing and machine learning.
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.