MDFULog: Multi-Feature Deep Fusion of Unstable Log Anomaly Detection Model

Round 1

Reviewer 1 Report

1. Clearly discuss the objective of your work in the abstract section.

2. Authors should cite some recent references of 2022 related to the work.

3. why did the authors select the given problem? discussion required in the introduction section.

4.  Advantages of the proposed work should be mentioned in a separate section.

5. Limitations of the proposed work should be modified in the conclusion section.

6. some discussion is required about the comparison to the other existing methods.

7. For the reader's groups some details for the figure are required.

 

 

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

1. There were numerous grammatical errors throughout the whole paper.   The authors are encouraged to carefully proofread the manuscript before resubmission.

2.  This paper was supposed to focus on anomaly detection.  But there were no discussion on the anomalies detected from the HDFS or OpenStack log.  These anomalies might not be fully captured by precision recall and F1 metrics as different anomalies might imply different severities.

3. Since the scalability is also of main concern, this paper should study the corresponding scalability of each algorithm being compared beyond accuracy vs log size.  Corresponding execution time vs. log size is also of great interests.

4. The comparison with other related approaches is quite reasonable.

5. Were correlated anomalies (including causally related) considered in the paper?  

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Accepted