ASAD: Adaptive Seasonality Anomaly Detection Algorithm under Intricate KPI Profiles

Round 1

Reviewer 1 Report

3. The abstract is too long, and the main aim of the research is not clear catchup by the reader. So, I recommend rewriter the abstract.
4. For readers to quickly catch the contribution in this work, it would be better to highlight major difficulties and challenges and your original achievements to overcome them more clearly in the abstract and introduction.
5. On page 2, the authors argue that “ This study develops a new eBeats clustering algorithm, which reduce the large time overhead of KPI sub-sequence clustering process. eBeats frst extracts the principal information based on discrete cosine transformation, then clusters the principal information”. I cannot find clear details about time reduction. So, please add more defiles in the results section about this experiment to prove the argument.

 

4. Having a table that summarizes the variables, sets, and notations will facilitate the reading of the paper. Please add a table with all the variables and sets used in the system model.
5. The majority of figures are not clear, and where are the y-axis and y-axis labels?
6. What is the computational complexity of the proposed solution? Although it is shown that the solution converges within a limited number of iterations, if the computational complexity of each iteration is high, this renders the solution to be computationally infeasible. Hence, the authors are encouraged to discuss the computational complexity of their proposed solution.
7. The language and presentation need to be enhanced. There are long sentences and some typing mistakes.
8. The discussion section in the present form is relatively weak and should be strengthened with more details and justifications.
9. The manuscript could be substantially improved by relying and citing more on recent literature.
10. It is interesting to make a comparison with other methods.
11. In the conclusion section, the limitations of this study suggested improvements in this work and future directions should be highlighted.
12. There are some references in the reference section with incomplete bibliographic information.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Please check the attached file.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

In this work, we present ASAD, a seasonal adaptive KPI anomaly detection algorithm. 
The paper needs further improvement to make ready for publication. I invite the reviewer to carefully address the following comments. 
Also, all the changes should be highlighted with a different color if a revision is submitted.
1)    The literature review of this article is very terse and most of the references are old. Many recent article discussing the anomaly detection and its application in different research fields are missed. Please discuss the following related works in Section 2 and add a Table that summarizes their characteristics, adopted ML models, their pros and cons, ..etc.
-    Graph neural network-based anomaly detection in multivariate time series
-    Abnormal electricity consumption detection based on ensemble learning
-    Artificial intelligence based anomaly detection of energy consumption in buildings: A review, current trends and new perspectives
-    Clustering-based anomaly detection in multivariate time series data
-    A novel approach for detecting anomalous energy consumption based on micro-moments and deep neural networks
-    Neural contextual anomaly detection for time series
-    Detection of appliance-level abnormal energy consumption in buildings using autoencoders and micro-moments 
-    Temporal convolutional autoencoder for unsupervised anomaly detection in time series
-    Smart power consumption abnormality detection in buildings using micromoments and improved K‐nearest neighbors
-    FluxEV: a fast and effective unsupervised framework for time-series anomaly detection
-    An anomaly detection framework for time series data: An interval-based approach
-    Learning graph structures with transformer for multivariate time series anomaly detection in iot
2)    The drawbacks and limitations of the proposed method should be highlighted in the Conclusion along with the future work.
3)    The paper needs a careful proofread as there some typos and grammatical issues.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

This paper investigates the seasonal adaptive KPI to anomaly detection. Also, a new eBeats clustering algorithm has been presented to reduce the overhead time of the KPI sub-sequence clustering process. The paper is well-organized and the idea is interesting. However, some minor modifications are required before publication. The abstract should be rewritten to better present the research gap and the achieved results of this study. Contributions of the paper are not clear, please provide the difference between this work and existing studies clearly. A section should be added to address the comparative analysis. The quality of figure 1 should be improved and please addressed the reference of the figure that is not from the authors' side. Please provide more description for figure 8. Please update the references to address more works published in recent years. For instance, the study in dynamic characteristics preserving data compressing algorithm for transactive energy management frameworks can be useful.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

See the attached pdf.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors amended all comments.

Reviewer 3 Report

The authors have carefully addressed all the comments. I have no further suggestions. The paper has significantly been improved.

Reviewer 4 Report

Comments have been addressed!

Reviewer 5 Report

The authors positively answered the major part of my concerns and, based on the changes applied, in my opinion the paper is ready for publication.