Differential Privacy-Based Spatial-Temporal Trajectory Clustering Scheme for LBSNs
Round 1
Reviewer 1 Report
This is a very interesting research paper. The proposed DP-STTC scheme provides a new and effective way to transform the location-based privacy protection mechanism into spatial-temporal trajectory protection. The idea and approach are original, and the experimental implementation demonstrates the novelty of the DP-STTC scheme. The research tackles the important privacy issue in the era of mobile computing and smartphone. The paper presented a scheme that could have a great application to protect personal privacy and personal data privacy from malicious use of personal location data collected by mobile carriers and mobile applications or platform owners. The presented research methodology is appropriate. The experimental evaluation is legitimate. However, there several issues could be addressed during the revision of the paper.
1. Section 3 is the key section; however, the description of each submission is not very clear. Even if the scheme is mathematical and abstract, the authors should give some simple examples to illustrate the abstract concept of mathematic expression to help readers better understand your idea and repeat or validate your work.
2. In scheme design, please provide more detailed information about the terminology, give more detail, and sample data of each parameter or variable, such as what is the condition? How is the condition number given or chosen? What are the spatiotemporal activity templates?
3. Figure 1 System Architecture is very vague.
3.1 “Actual Location Data”, does it mean the dataset you are using, or the mobile device’s location data collected and stored?
3.2 “Mobile Client, is it a mobile application running the DP-STTC scheme on the mobile device?
3.3 “Server”, what is the server?
The data flow looks not right. Is Disturbed Trajectory Data the output? There is no explanation about how the architecture works and how it can protect the location data from the “attackers”.
4. Figure 2, “LPPM”, I guess it is “Location Privacy Protection Mechanism”, no explanation about it.
5. Eq. (6): it is correct? Pr(A | EVENT) = Pr(A, EVENT)/Pr(EVENT)? Please explain!
6. Eq. (7) – (13) need explanations to connect the quadratic programming method with your application. Please give some examples. What do the variables or parameters, a b, and c mean and represent in your application?
7. The two datasets, GeoLife and Beijing POI datasets, how are they used together in your experimental evaluation, and why?
There are some general issues, please address them in the revision:
1. In Sections 3 and 4 where the DP-STTC scheme is presented and described, there is no mention of DP-STTC at all in the two sections, why?
2. Many references, 22 out of 32 are from one country, including three of the authors’ own, which looks not normal.
For the above issues, it may be because I don’t understand your paper correctly. However, if I can’t understand your paper properly, there is a problem with the paper presentation. It is only the purpose for me to raise the above issues or concerns that I hope this valuable research can be published and can contribute to the literature for other researchers and to protect personal privacy information for people.
Author Response
Dear reviewer,
The attachment is the response to the comments.
Thank you very much!
Kind regards
Liang Zhu
Author Response File: 
Author Response.pdf
Reviewer 2 Report
1.
The statement "\delta-Location Set [22]" on Line 140 is incorrect. It should read "\delta-Location Set [18]."
2.
Many symbols are undefined, making it difficult to comprehend the proposed algorithm. Below are some examples. Please provide clear definitions for them:
Symbols i, l, and Z on Line 140.
Symbol c on Line 144.
Symbol m on Line 175. It's worth noting that the authors use symbol m to denote the size of the location domain on Line 160. However, the symbol m on Line 175 has a different implication. The authors should therefore use a different symbol on Line 175 and provide its description.
Symbol M on Line 244.
3.
Definitions 6 and 7 have been previously introduced in the following article:
[*1] Cao, Yang, et al. "PriSTE: from location privacy to spatiotemporal event privacy." 2019 IEEE 35th International Conference on Data Engineering (ICDE). IEEE, 2019.
The authors need to provide appropriate citation for this reference.
4.
Could you elucidate the difference between l_t and l_T as mentioned on Lines 183 and 184?
5.
Equation (5) on Line 247 was previously presented in [18]. As such, Section 4.1 doesn't offer any novel technical contributions.
Furthermore, there appear to be errors in Equation (5). The authors are advised to revisit [18] and ensure they fully comprehend its algorithm.
6.
Algorithm 1 in this manuscript is identical to Algorithm 1 introduced in [*1]. Plagiarism is strictly prohibited.
7.
Equations 8-13 were proposed in the following article [*2].
[*2] Cao, Yang, et al. "Protecting spatiotemporal event privacy in continuous location-based services." IEEE Transactions on Knowledge and Data Engineering 33.8 (2019): 3141-3154.
8.
Algorithm 3 in this manuscript is identical to Algorithm 3 introduced in [8].
Overall, the proposed algorithm in this manuscript heavily leans on [8,18,*1,*2]. However, it seems the authors might not have fully grasped the algorithms from [8,18,*1,*2], leading to multiple inaccuracies in both the proposed algorithm and the manuscript's statements.
English very difficult to understand.
Author Response
Dear reviewer,
The attachment is the response to the comments.
Thank you very much!
Kind regards
Liang Zhu
Author Response File: Author Response.pdf
Reviewer 3 Report
In the work titled “Differential Privacy Based Spatial-Temporal Trajectory Clustering Scheme over LBSNs”, the authors propose a Differential Privacy-based Spatial-Temporal Trajectory Clustering (DP-STTC) scheme, claiming that such a scheme can better protect user privacy, achieving also better accuracy of trajectory clustering, with regard to the state-of-the-art solutions.
The "Introduction" and "Related Works" sections of the paper do not provide a comprehensive overview of the research area under consideration. It is recommended to expand these sections to better encapsulate the broader landscape of the research domain. In particular, the "Introduction" section should be utilized to offer a high-level description of the field, setting the stage for further exploration in the "Related Works" section, which should be renamed as “Background and Related Work”.
The references do not appear very updated, so the authors should check if there are more recent works among those they mentioned and, in any case, accordingly to my previous observation, they should add and discuss additional works very close or directly related to the research field taken into account, such as, just by way of example:
(-) Kim, Tai-Hoon, et al. "A novel trust evaluation process for secure localization using a decentralized blockchain in wireless sensor networks." IEEE access 7 (2019): 184133-184144.
(-) Saia, Roberto, et al. "A blockchain-based distributed paradigm to secure localization services." Sensors 21.20 (2021): 6814.
(-) Shi, Xiufang, et al. "Resilient privacy-preserving distributed localization against dishonest nodes in Internet of Things." IEEE Internet of Things Journal 7.9 (2020): 9214-9223.
(-) Jiang, Hongbo, et al. "Location privacy-preserving mechanisms in location-based services: A comprehensive survey." ACM Computing Surveys (CSUR) 54.1 (2021): 1-36.
(-) … And so on!
In addition, the authors should check that citations are correct, as, for example, the following citation is repeated twice (i.e., [7] and [27]):
Qiao, D.; Liang, Y.; Ma, C.; Zhang ,H. Semantic Trajectory Clustering via Improved Label Propagation With Core Structure. IEEE Sens J. 2022, 22, 639-650.
About the “Conclusions and Future Work”, the authors should expand the given information by recapping, in brief, all the main steps of their manuscript, in order to offer a brief but complete summary of the work carried out to the readers.
In this section, the authors should also better underline the possible advantages of the proposed approach with regard to both the state-of-the-art solutions and the real-world applications.
The formalization of the proposed scheme and the conducted experiments are clearly exposed. However, the authors should offer additional details about their selection process for the datasets and the competitors they considered for performance evaluation.
Without going into the details of the formal and grammatical problems of the proposed manuscript, I suggest to re-read carefully to remove some typos/errors and rewrite some sentences more clearly, e.g.:
- “services in future” instead of “services in the future”;
- “user privacy, but also” instead of “user privacy but also”;
- “or regularly traveling between” instead of “or regularly travels between”;
- “three aspects as following.” instead of “three aspects as follows.”;
- “performance of proposed DP-STTC” instead of “performance of the proposed DP-STTC”;
… And so on, considering that I recognized all the above mistakes/problems in the first pages of the manuscript”;
Author Response
Dear reviewer,
The attachment is the response to the comments.
Thank you very much!
Kind regards！
Liang Zhu
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The revised manuscript shows improvement over the previous version; however, I still find the overview of the algorithm unclear. Could you please provide more details on "LPPM" as described in Algorithm 1? Additionally, you mention "adjusting privacy parameters and generate L_t^*" but do not elaborate on this aspect. Could you clarify what this entails?
I also recommend improving Algorithm 1 and providing specific examples. Offering concrete examples will enable me to provide more targeted feedback.
Moreover, please prove that the proposed algorithm ensures differential privacy.
Moderate editing of English language required
Author Response
Dear Reviewer,
The attachment is the responses to the comments.
Special thanks to you for your good comments.
Best regards,
Liang Zhu
Author Response File: Author Response.pdf
Reviewer 3 Report
The revised version of the manuscript has been improved compared to the first submission of the authors and all the problems I highlighted in my review have been solved.
Author Response
Dear Reviewer,
Special thanks to you for your good comments.
Best regards,
Liang Zhu
Round 3
Reviewer 2 Report
Since the careful response to the peer review comments was confirmed, I make a decision to accept the manuscript.
Minor editing of English language required
