Sediment Level Prediction of a Combined Sewer System Using Spatial Features
, Ramon Bejar 2, Edgar Rubión 1, Lluís Echeverria 1, Francisco Javier Varela Alegre 3 and Lluís Corominas 4Round 1
Reviewer 1 Report
The authors used different machine learning to reduce the number of cleaning the combined sewer system, using the regression model to predict the sediment level, using the short-term regression model to predict the imminent sediment level, and using the classification model to identify whether the section needs cleaning. Besides, the conclusion offers promising future topics. I only have some questions for the authors.
- Why is the R2 score of 0.76 good enough for this study?
- The recall is lower than other studies [20, 22], is the imbalance? How to improve this problem?
- Will the comparisons of speed and size of models be necessary for this study?
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The authors have compared three different approaches and machine learning methods used for predicting sewer inspection requirements. The manuscript explicitly highlights the gaps while citing relevant references. I have the following comments:
- Some aspects of dataset are missing from the manuscript. E.g. when authors say 34000 inhabitants does it implies household or residents since there can be multiple residents in one household.
- How many datapoints were available for training? Was the data enough to avoid overfitting?
- Was it confirmed that there was no overfitting? For reference, check the following paper using ANN and discussing overfitting:
Vyas, Shruti, Subhabrata Das, and Yen-Peng Ting. "Predictive modeling and response analysis of spent catalyst bioleaching using artificial neural network." Bioresource Technology Reports 9 (2020): 100389. - Explanation or discussion on why some method performed better in an approach and not in other. e.g ANN performs better in the first and second approach however not in the third approach.
- Figure 1, x-axis values of first three columns do not match with text.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
Please find in the attachment.
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The authors have answered most of my questions.
