Semi-Supervised Adversarial Transfer Networks for Cross-Domain Intelligent Fault Diagnosis of Rolling Bearings

Round 1

Reviewer 1 Report

1.The research question is clearly stated but a revision in evaluation of the proposed method is required.

2.The results need to be presented more clearly on how they are extracted from the source data.

3.The explaining of the initial steps of the process of collecting vibration signals to label generation of the images have been repeated throughout the paper.

4.Equation (10) on page 8 does not match with the one proposed in the stated reference [51]

5.Figure 3 on page 7, Stage 1 uses the same equation (10) with incorrect data, and also proper labeling of the image is required.

6.Proper explanation of figure 10 is required.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

1. Similarity is more than 30% without references.

2. Manuscript should be improved for its English, grammar, and content flow.

3. Why time domain data is converted into an image dataset? Justify this with their advantages and limitations for the dataset used in this paper.

4. The literature is divided into different sub-section, which is appreciable. Summarize/ present the same literature in tabular form for better clarity. 

5. Section 3.1 - The effect of assumption on the results and their interpretation should be clearly mentioned.

6. All equations should be checked carefully.

7. The figures should be prepared as per the guidelines of the journal.

8. What information is lost while converting the time domain data into image data? How is this problem to be overcome?

9. Figure 12 - the critical observations drawn from the table should be elaborated.

10. Conclusion should be updated with more critical observations and conclusions.

Author Response

Please see the attachment.

Author Response File: Author Response.docx