A CNN-Based Method for Heavy-Metal Ion Detection

Round 1

Reviewer 1 Report

Please see the attachment

 

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

1.       Has this new method validated against Golden standards?  Comparison of results may be shown in discussion / conclusion part.

2.       Why specifically only 1200 samples were created? Basis for this number?

3.       What Preprocessing methods used to avoid loss of information from data acquired? Detail?

4.       Provide Table with details of features extracted. How do you optimize the number of features?

5.       70% data used for Training. These data samples should be the best one to get good accuracy. How do you select this? Basis?

6.       Figure 4 – How do you arrive at NN diagram with input layer:  800, hidden: 64 and output 3 neurons. Design parameters may be explained.

7.       Include Flowchart / algorithm / steps to train and test CNN

8.       Accuracy is almost 100%. Need comparison with conventional methods to justify.

9.  Only accuracy shown. Author could have tried for few misclassification detections with different concentrations. This is essential to ensure perfect design of the network. Specificity and Sensitivity parameters not discussed.

1.   Author could have tried with other computational methods such as SVM and compared with CNN. Any specific reasons for not selecting?

1.   Few of the References not discussing about computational methods. Seems not relevant.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf