IoT-Enabled Chlorine Level Assessment and Prediction in Water Monitoring System Using Machine Learning
, Surya Kalyan Chakravarthy Nidamanuri 3,†, S. Jayalakshmi 4,†, Ramadevi Bogani 5, P. Thanapal 6,†, Shahada Syed 7, Boppudi Venkateswarlu 4 and Jafar Ali Ibrahim Syed Masood 1,*,†
Round 1
Reviewer 1 Report
First, the paper lacks sufficient detail on the machine learning algorithms used. While the authors briefly mention using "machine learning techniques," they do not provide any further information on which algorithms were employed or how they were implemented. This makes it difficult for readers to fully understand and evaluate the proposed approach.
In addition, the system design presented in the paper is relatively trivial and does not seem to offer any significant innovations or improvements over existing solutions. The authors should consider including more detailed descriptions of the system architecture and how it addresses the challenges of chlorine level assessment and prediction in water monitoring systems.
Finally, the scientific contribution of the paper is not clear. While the authors present some preliminary results, they do not provide sufficient analysis or discussion to demonstrate the significance of their findings or how they contribute to the field.
Overall, the paper would benefit from more thorough development and clarification of the proposed approach, as well as a stronger focus on the scientific contributions and implications of the work. I recommend revising the paper to address these issues before resubmitting for review.
Author Response
We appreciate the valuable suggestions from the reviewer, which are greatly helpful for improving this manuscript.
- Comments
First, the paper lacks sufficient detail on the machine learning algorithms used. While the authors briefly mention using "machine learning techniques," they do not provide further information on which algorithms were employed or how they were implemented. This makes it difficult for readers to fully understand and evaluate the proposed approach
Response:
Sorry for the inconvenience caused; as per reviewer comments, we detailed the machine learning techniques and the specific algorithm, which is the fuzzy rule in the manuscript.
“Fuzzy logic can be used to predict chlorine levels in the water. Fuzzy rules describe the relationship between input variables (e.g., temperature, pH, etc.) and the output variable (chlorine level). These rules can be combined to form a fuzzy inference system, making predictions based on input data. The accuracy of the predictions depends on the quality of the fuzzy rules and the accuracy of the input data. It is important to validate the fuzzy inference system using actual data to ensure accuracy.
Comments
In addition, the system design presented in the paper is relatively trivial and does not seem to offer any significant innovations or improvements over existing solutions. The authors should consider including more detailed descriptions of the system architecture and how it addresses the challenges of chlorine level assessment and prediction in water monitoring systems.
Response:
Sorry for the inconvenience caused; as per reviewer comments, we detailed the challenges of chlorine level assessment and prediction in water monitoring systems.
“There are several challenges that need to be addressed when using the proposed model to predict chlorine levels in water monitoring systems, including the data quality and availability. The machine learning algorithms require large amounts of accurate data to train the model. Ensuring that the data collected is high quality and available in sufficient quantities is crucial for developing an accurate prediction model. Furthermore, feature selection is important to identify the key factors that affect chlorine levels in the water, as these will become the input variables for the prediction model. Careful selection of the input variables is crucial for obtaining accurate predictions. Additionally, a simple model may not capture the complex relationships between the input variables and chlorine levels. At the same time, too complex a model may overfit the data, leading to poor predictions for new data. Finally, model evaluation is important in evaluating the model’s performance using appropriate metrics such as mean absolute error, and root means square error or R-squared. This helps to determine the predictions’ accuracy and identify areas for improvement.”
Comments
Finally, the scientific contribution of the paper is not clear. While the authors present some preliminary results, they do not provide sufficient analysis or discussion to demonstrate the significance of their findings or how they contribute to the field.
Response
Thank you for the valuable suggestions. As per the reviewer’s suggestions, we have included the purpose of the proposed model validation.
“Model validation is essential in water monitoring systems' chlorine level assessment and prediction. The purpose of model validation is to assess the accuracy of the model's predictions by comparing its predictions with the actual chlorine levels in the water. As mentioned above, the most common model validation approach divides the data into training and testing datasets. The training dataset is used to train the proposed algorithm, while the testing dataset is used to validate the model’s performance. The model is trained on the training dataset, and its predictions are compared with the actual chlorine levels in the testing dataset. This comparison calculates metrics such as recall, precision, and F-score. These metrics give a quantitative measure of the accuracy of the predictions and can be used to compare the performance of the proposed model.”
Reviewer 2 Report
The theme of the article is quite pertinent, however it is important to take into account the following considerations:
1. Reviewing the body of the article, the implementation of the decision tree algorithm is noted, however, it is not explained in the abstract. The contribution in terms of generating new knowledge must be clearly defined.
2. It seems that this text is part of the introduction and is in related works.
The remaining sections are categorized as follows. In Section 2, we briefly cover some pertinent literature. Section 3 describes the recommended methodology, Water Quality Analysis Framework, for measuring and analyzing the current pollution level and estimating its future status. In Section 4, the experimental findings and comments are detailed. Section 7 concludes with the conclusion.
3. It is important that the review of the work carried out allows, among other things, to be able to show the contribution of the research, compared to other developments that are already found in the literature, I believe this section should be expanded a little more.
Allow me to suggest some references that could enrich the state of the art of the article:
- Ariza-Colpas, P. P., Ayala-Mantilla, C. E., Shaheen, Q., Piñeres-Melo, M. A., Villate-Daza, D. A., Morales-Ortega, R. C., ... & Afzal, M. (2021). SISME, estuarine monitoring system based on IOT and machine learning for the detection of salt wedge in aquifers: case study of the Magdalena River estuary. Sensors, 21(7), 2374.
- Huang, R., Ma, C., Ma, J., Huangfu, X., & He, Q. (2021). Machine learning in natural and engineered water systems. Water Research, 205, 117666.
- Zhu, M., Wang, J., Yang, X., Zhang, Y., Zhang, L., Ren, H., ... & Ye, L. (2022). A review of the application of machine learning in water quality evaluation. Eco-Environment & Health.
- Virnodkar, S. S., Pachghare, V. K., Patil, V. C., & Jha, S. K. (2020). Remote sensing and machine learning for crop water stress determination in various crops: a critical review. Precision Agriculture, 21(5), 1121-1155.
4. The reference to figures 3 and 4 cannot be seen in the manuscript.
5. In the description of the selection of characteristics, the development of each of the phases that were carried out in the investigation should be expanded.
6. To enrich and strengthen the proposed results, it would be important to include other algorithms in the experimentation in order to determine why this is the best solution. Why select just these algorithms? It is a good question to which the article must give an answer in order to determine the novelty of the experimentation.
7. The conclusion should be strengthened based on the impact on the community or the model generated.
8. It should be added which are the future works of the investigation.
Author Response
- Comments
It seems that this text is part of the introduction and is in related works.
The remaining sections are categorized as follows. In Section 2, we briefly cover some pertinent literature. Section 3 describes the recommended methodology, Water Quality Analysis Framework, for measuring and analyzing the current pollution level and estimating its future status. In Section 4, the experimental findings and comments are detailed. Section 7 concludes with the conclusion.
Response:
Sorry for the inconvenience caused. As per the reviewer’s comments, we completely modified the content mentioned below.
“The remaining sections are categorized as follows. In Section 2, we briefly cover some pertinent literature. Section 3 describes the proposed methodology, Chlorine level Assessment and Prediction in Water Monitoring System using a fuzzy set, specifically using a decision tree algorithm. In section 4, the data processing workflow of the proposed model is detailed. In Section 5, the results and discussion are illustrated. Section 7 concludes with the conclusion.”
Comments
It is important that the review of the work carried out allows, among other things, to be able to show the contribution of the research, compared to other developments that are already found in the literature, I believe this section should be expanded a little more.
Allow me to suggest some references that could enrich the state of the art of the article:
- Ariza-Colpas, P. P., Ayala-Mantilla, C. E., Shaheen, Q., Piñeres-Melo, M. A., Villate-Daza, D. A., Morales-Ortega, R. C., ... & Afzal, M. (2021). SISME, estuarine monitoring system based on IOT and machine learning for the detection of salt wedge in aquifers: case study of the Magdalena River estuary. Sensors, 21(7), 2374.
- Huang, R., Ma, C., Ma, J., Huangfu, X., & He, Q. (2021). Machine learning in natural and engineered water systems. Water Research, 205, 117666.
- Zhu, M., Wang, J., Yang, X., Zhang, Y., Zhang, L., Ren, H., ... & Ye, L. (2022). A review of the application of machine learning in water quality evaluation. Eco-Environment & Health.
- Virnodkar, S. S., Pachghare, V. K., Patil, V. C., & Jha, S. K. (2020). Remote sensing and machine learning for crop water stress determination in various crops: a critical review. Precision Agriculture, 21(5), 1121-1155.
Response:
We have included the above-mentioned reference in the article per reviewer suggestions.
Comments
The reference to figures 3 and 4 cannot be seen in the manuscript.
Response:
Sorry for the inconvenience caused. As per reviewer comments, we cited figure 3 and figure 4 in the manuscript.
Comments
In the description of the selection of characteristics, the development of each of the phases that were carried out in the investigation should be expanded.
Response:
- Thank you for the valuable suggestions. As per reviewer suggestions, we have included the selection of characteristics and gave justification for each of the phases
“Temperature can affect the solubility of the chlorine compounds in the water and the reaction kinetics between the chlorine and other substances in the water. For example, warmer temperatures can increase the solubility of chlorine compounds and can speed up the reaction kinetics, leading to a higher chlorine aggregate ratio. It is important to monitor the groundwater's temperature and consider it when assessing the chlorine aggregate ratio. This can help ensure that accurate predictions are made and the water quality is properly maintained.”
- Comments
To enrich and strengthen the proposed results, it would be important to include other algorithms in the experimentation in order to determine why this is the best solution. Why select just these algorithms? It is a good question to which the article must give an answer in order to determine the novelty of the experimentation.
Response:
Thank you for the valuable suggestions. As per reviewer suggestions, we have included justification for decision trees in the proposed model.
- The proposed model uses a decision tree algorithm, an intuitive and interpretable method for making predictions. It can also handle missing data and is not sensitive to outliers. However, it can become complex and overfit the data if fewer branches and input variables exist. To address this issue, decision trees can be pruned by removing branches that do not significantly improve the accuracy of the predictions. The pruning process helps reduce the model’s complexity and improve its accuracy. The decision tree algorithm is a popular choice for chlorine level assessment and prediction in water monitoring systems due to its ease of use and ability to handle complex data.
Comments
The conclusion should be strengthened based on the impact on the community or the model generated.
Response:
We have modified the conclusion and updated the manuscript per the reviewer's suggestion.
“The chlorine level must be monitored regularly to ensure that it is within safe limits, neither too low to be effective nor too high to cause health problems. Accurate assessment and prediction of chlorine levels in water monitoring systems are essential for controlling and maintaining the water treatment process, detecting and addressing any deviations in water quality, and ensuring compliance with regulations and standards. Regular monitoring of chlorine levels also helps water utilities to optimize their water treatment processes and reduce costs associated with excess chlorine addition.”
Reviewer 3 Report
this paper deals with level assesment of chlorine using IOT and machine learning. the paper is well structured. however, I have some remarks:
1- Chlorine is a widely used natural water disinfectant, but because it combines with ammonia's nitrogen to generate chloramines, the measurement of free chlorine is not as accurate as it should be. The authors did not address this issue. Could you please explain this problem? in the work presented in this paper, have you considered this problem?
2- the authors should spell out the full term of the abbreviations used in the papers such as SVM and KNN, for example.
3- can the authors explain the idea behind choosing these methods (SVM and KNN) instead of other methods published in the literature such as kernel extreme learning machine (KELM) and support vector regression (SVR).
Author Response
Comments
Chlorine is a widely used natural water disinfectant, but because it combines with ammonia's nitrogen to generate chloramines, the measurement of free chlorine is not as accurate as it should be. The authors did not address this issue. Could you please explain this problem? in the work presented in this paper, have you considered this problem?
Response:
This manuscript highlights the accurate assessment and prediction of chlorine levels in water monitoring systems which is essential for controlling and maintaining the water treatment process. In the upcoming manuscript, as per the reviewer’s suggestion, we will incorporate the measurement of free chlorine level.
Comments
The authors should spell out the full term of the abbreviations used in the papers such as SVM and KNN, for example.
Response:
As per the reviewer’s suggestion, we have included the abbreviations for SVM and KNN in the manuscript.
Comments
Can the authors explain the idea behind choosing these methods (SVM and KNN) instead of other methods published in the literature such as kernel extreme learning machine (KELM) and support vector regression (SVR).
Response:
Thank you for the valuable suggestion. SVM and KNN are used just for comparison with the proposed model. The proposed model uses Fuzzy rules to predict and assess chlorine level.
Round 2
Reviewer 2 Report
In the new version of the article, the suggested revisions have been contemplated
