A Survey on Change Detection and Time Series Analysis with Applications
Round 1
Reviewer 1 Report
Dear authors,
Thank you very much for your efforts writing this paper. The academic writing style is clear and precise. The content is logically arranged, well founded and comprehensible in its reasoning. The literature used is up to date and relevant for treating the topic of research.
Authors identified the lack of review papers about most recent methods that can rigorously analyze non-stationary time series, therefore this manuscript is a review focused on many traditional and recent techniques for time series analysis and change detection, including wavelet and spectral analyses with their advantages and weaknesses. The topic of the paper is interesting and relevant.
The illustrations, tables, figures and graphs are shown properly, adding very relevant information.
Although in my opinion this manuscript could be directly accepted to publication in present form, please take into consideration the following minor comments to improve your manuscript quality:
C1. Please include your findings and conclusion in the last paragraph of the abstract.
C2. Please consider adding research question/s and hypothesis/es at the end of section 1.
C3. It is recommended to explain the manuscript organization at the end of section 1.
C4. Please expand conclusions section and discuss your findings.
C5. Please consider adding the limitations of this investigation.
I congratulate the authors for this interesting investigation and wish them the most success in their research activities.
Thank you very much for your efforts and for your valuable scientific contribution.
Author Response
Authors’ Responses to the comments of Reviewer 1.
Dear Reviewer,
We would like to thank you very much for providing us these very useful comments that significantly helped us to improve our paper. Please kindly see below our responses to your comments that are also highlighted in the manuscript.
Dear authors,
Thank you very much for your efforts writing this paper. The academic writing style is clear and precise. The content is logically arranged, well founded and comprehensible in its reasoning. The literature used is up to date and relevant for treating the topic of research.
Authors identified the lack of review papers about most recent methods that can rigorously analyze non-stationary time series, therefore this manuscript is a review focused on many traditional and recent techniques for time series analysis and change detection, including wavelet and spectral analyses with their advantages and weaknesses. The topic of the paper is interesting and relevant.
The illustrations, tables, figures, and graphs are shown properly, adding very relevant information.
Although in my opinion this manuscript could be directly accepted to publication in present form, please take into consideration the following minor comments to improve your manuscript quality:
C1. Please include your findings and conclusion in the last paragraph of the abstract.
Response. We have added the following sentence at the end of Abstract: Understanding of the methods presented herein is worthwhile to further develop and apply them for unraveling our universe.
C2. Please consider adding research question/s and hypothesis/es at the end of section 1.
Response. We have added the following texts:
Herein, we discuss how one may: 1) extract the useful information from a time series theoretically and empirically, 2) attenuate noise and regularize time series, 3) detect and classify changes in the time series components, and 4) analyze unequally spaced time series without any interpolations while considering the observational uncertainties.
C3. It is recommended to explain the manuscript organization at the end of section 1.
Response. We have modified some of the texts as follows:
This paper starts by reviewing the most popular and traditional time series analysis methods and continues by reviewing some of the most recent methods of analyzing non-stationary and unequally spaced time series. More specifically:
In Section 2, several popular decomposition methods into frequency domain are briefly reviewed, such as the Fourier transform and least-squares spectral analysis and their modifications.
In Section 3, several popular time-frequency decomposition methods are discussed, including the short-time Fourier and wavelet transforms, Hilbert-Huang transform, constrained least-squares spectral analysis, and least-squares wavelet analysis, and then two methods of analyzing multiple time series together are reviewed.
In Section 4, several change or breakpoint detection methods within non-stationary time series are studied.
In Section 5, many applications of the methods in applied sciences along with other techniques of time series analysis are briefly mentioned.
Finally, the conclusion, findings, and limitations of this investigation are briefly summarized in Section 6.
C4. Please expand the conclusions section and discuss your findings. And C5. Please consider adding the limitations of this investigation.
Response. We have expanded the Conclusions Section as follows:
In this paper, many frequency and time-frequency decomposition methods via Fourier and least-squares analyses were studied. Many change detection and monitoring methods were also briefly reviewed, and some of the applications of all methods were listed in Table 1. There are many more robust time series analysis techniques proposed by researchers that were not mentioned here. In many practical applications, time series contain seasonal and trend components. Simultaneous estimation of the statistically significant components can provide more accurate and reliable estimates for the time series components and so more appropriate for change detection and monitoring. The time series components can also be estimated more accurately when considering the observational uncertainties. Therefore, the observations with higher uncertainties get less weight during the analysis and vice versa. Computational complexity optimization is another major challenge in time series analysis when dealing with big data sets. An inappropriate algorithm modification for reducing the computational cost can produce unreliable and inaccurate results. Each method presented herein has advantages and weaknesses, and there is plenty of room for researchers and scientists to expand and improve the existing methods.
I congratulate the authors for this interesting investigation and wish them the most success in their research activities.
Thank you very much for your efforts and for your valuable scientific contribution.
We thank you very much for these valuable comments
Best regards,
Ebrahim Ghaderpour
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper presents a solid overview of time series analysis and can be published after minor spelling corrections.
Author Response
Authors’ Responses to the comments of Reviewer 2.
The paper presents a solid overview of time series analysis and can be published after minor spelling corrections.
Dear Reviewer,
We would like to thank you very much for reviewing our manuscript. We have double-checked the paper for any typos and/or grammar issues.
Sincerely yours,
Ebrahim Ghaderpour
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors provided a review work as a summary of quite a few frequency and time/frequency decomposition methods vis frequency and least-squares analysis. The references cover classic signal processing algorithms and recent advances. The topic is useful to researchers in this field. This work could be a good introduction to the readers as an overview of existing work and guidence of the advantages of different methods. Overall the paper is well-organized. Most figures are presented with good quality. A few suggestions which may further improve the manuscript. 1. Line 3 and 672, is it a typo for "analyses", should be "analysis"? 2. In Fig. 1 and 3, the y-axis in a few subplots are without a meaningful unit? Is it possible to apply some standard units to them?Author Response
Authors’ Responses to the comments of Reviewer 3.
Dear Reviewer,
We would like to thank you very much for reviewing our manuscript and for your valuable comments. Please kindly see below our responses to your comments that are also highlighted in the manuscript.
The authors provided a review work as a summary of quite a few frequency and time/frequency decomposition methods vis frequency and least-squares analysis. The references cover classic signal processing algorithms and recent advances. The topic is useful to researchers in this field. This work could be a good introduction to the readers as an overview of existing work and guidance of the advantages of different methods. Overall, the paper is well-organized. Most figures are presented with good quality. A few suggestions which may further improve the manuscript.
- Line 3 and 672, is it a typo for "analyses", should be "analysis"?
Response. We have checked these phrases to ensure their correctness. The word “analyses” is the plural of “analysis”. Thus, instead of writing “wavelet analysis and spectral analysis”, we simply say “wavelet and spectral analyses”, etc.
- In Fig. 1 and 3, the y-axis in a few subplots are without a meaningful unit? Is it possible to apply some standard units to them?
Response. We appreciate your comment. The y-axis value can have any physical unit as the time series are simulated. To address your comment; however, we have added the following sentence on page 5 lines 170-171:
For example, f(t) may be an ambient temperature in degree Celsius (C) at time t, and so the amplitude will be in C in Figure 1.
We hope these changes are satisfactory and thank you again for reviewing our manuscript.
Best regards,
Ebrahim Ghaderpour
Author Response File: Author Response.pdf
