Unlocking the Potential of Soft Computing for Predicting Lubricant Elemental Spectroscopy
Round 1
Reviewer 1 Report
Review Title: An Insightful Approach to Spectroscopy Prediction of Lubricant Components by Soft Calculation The article "Unlocking the Potential of Soft Computing for Predicting Lubricant Elemental Spectroscopy" represents a groundbreaking study on predictive maintenance of mechanical systems. This study, based on the electrical properties of oil samples, evaluates the effectiveness of soft computing models in predicting spectroscopy of engine lubricant components. Positive aspects of the article: The authors focused on the important area of ​​predictive maintenance, predicting the chemical composition of lubricants, which has a direct impact on their performance and equipment longevity. This is an innovative approach to the problem. The study uses multiple soft computing models (RBF, ANFIS, SVM, MLP, GPR) to compare their performance in predicting element spectroscopy. This provides readers with an understanding of which models are most effective in the context of the problem under study. Application error metrics such as MAPE, RMSE, and EF provide a solid basis for evaluating the performance of individual models. Sensitivity analyzes performed reveal that the best results are obtained using all three electrical properties as inputs. This important insight may contribute to improving predictive models in the future. The authors admit that despite the limitations, their work provides a solid basis for further research. Such a self-critical approach is essential to true science. Recommendations for improvement: The study used a small data set (49 lubricant samples), which may have affected its generalization. Future studies could use a larger data set to make the results more reliable. The authors point out that the practice of measuring the electrical properties of lubricants and devices is not common in their country, leading to a limited number of samples. Possibility of cooperation with institutions from other countries, where such measurements are more common common, could help to gather more data. While the authors note the need for further research, they could suggest more specifics. For example, what other machine learning models can be used? What other optimization techniques can improve results? While the article provides a practical framework for predicting the condition of lubricants based on electrical measurements, more information on potential implementation in real-world monitoring systems would be useful. In conclusion, the paper represents a significant step forward in the field of predictive maintenance, offering valuable insights into the use of soft computing to predict the chemical composition of motor lubricants. Despite some limitations, this work provides a solid basis for further research.
Please supplement the article with items: https://doi.org/10.1177/1350650120964026, https://doi.org/10.3390/ma16114092, https://doi.org/10.3390/nano13050955, https://doi.org/ 10.1016/j.energy.2022.126002, https://doi.org/10.3390/s23135970 Then write conclusions in points. Emphasize the innovation of the topic more. I don't see any significant news. The article is good but please advertise it more.
Author Response
The response of the esteemed reviewer can be viewed in the attached file, adhering to the suggested format.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript scope is dealing with the research on utilization of different computing models for predictive lubricant elemental spectroscopy to improve condition monitoring based maintenance. The paper is well written and organized. I recommend it to publish after minor revision as the following:
1. Emphasize novelty aspect of this research at the introduction.
2. There are some typo errors (337-347 lines). Please polish the manuscript carefully.
3. I do not see is there considered content of water in the engine oil, because it can significantly influence dielectric property. If the water content is on stable level, presented results could be valid.
4. Back up the conclusion section with numerical data.
Author Response
The response of the esteemed reviewer can be viewed in the attached file, adhering to the suggested format.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Thank you very much for clarification and accepted suggestions.
