ICACT20230212 Question.1
Questioner: lp_saikia@yahoo.co.in 2023-02-16 오후 8:37:39 |
ICACT20230212 Answer.1
Answer by Auhor 48838968@qq.com 2023-02-16 오후 8:37:39
Chrome  Click!! |
| Why your clustering algorithm based on binary tree? |
Clustering the data in the original data set can effectively reduce the loss of data information and improve the availability of data when similar data are divided together and then processed anonymously. |
ICACT20230212 Question.10
Questioner: M11007513@gapps.ntust.edu.tw 2023-02-19 오후 5:38:16 |
ICACT20230212 Answer.10
Answer by Auhor 48838968@qq.com 2023-02-19 오후 5:38:16
Chrome Click!! |
| According to the last mentioned problems and challenges, is there any possible direction to solve them? |
|
ICACT20230212 Question.2
Questioner: lp_saikia@yahoo.co.in 2023-02-16 오후 8:40:15 |
ICACT20230212 Answer.2
Answer by Auhor 48838968@qq.com 2023-02-16 오후 8:40:15
Chrome Click!! |
| Your proposed binary tree-based clustering algorithm BTCA is compared with the k-anonymous classical algorithm KACA algorithm, please point out findings. |
After taking different K values, the experiment shows that the BTCA algorithm proposed in this paper has less information loss than KCAC algorithm. And with the increase of K value, this advantage is more obvious and data availability is higher. |
ICACT20230212 Question.3
Questioner: lp_saikia@yahoo.co.in 2023-02-16 오후 8:41:40 |
ICACT20230212 Answer.3
Answer by Auhor 48838968@qq.com 2023-02-16 오후 8:41:40
Chrome Click!! |
| What are the limitations of your work? |
In this paper, only a single sensitive attribute in the dataset is selected during the experiment. When the dataset contains high-dimensional sensitive attributes, the efficiency and availability of the algorithm may be insufficient. Moreover, the biggest challenge of integrating the differential privacy mechanism is the balance between privacy and availability. How to reduce noise while maintaining the privacy of the result data, adjust the data privacy method based on the trade-off between utility and privacy, and create indicators to evaluate and measure the quality of model information loss and privacy protection is a challenge for future research. |
ICACT20230212 Question.4
Questioner: nvlinh@cs.ccu.edu.tw 2023-02-17 오전 11:47:32 |
ICACT20230212 Answer.4
Answer by Auhor 48838968@qq.com 2023-02-17 오전 11:47:32
Chrome Click!! |
| What are your contributions? |
With the acceleration of the process of data opening and sharing in the power industry, the risk of sensitive data leakage it faces is also gradually increasing. Privacy protection is one of the hot issues in the research of privacy disclosure control technology in data publishing at present, and k-anonymity is the hotspot of privacy protection research in recent years. In this paper, a hierarchical DP-K anonymous data publishing model based on binary tree clustering is proposed to solve the problem that the existing k-anonymity schemes lack the consideration of attribute sensitivity and minimize the amount of information loss. A clustering algorithm based on binary tree (BTCA) is proposed to divide similar data records into the same equivalent class, which can improve the clustering effect, reduce the information loss caused by the publication of anonymous data sets, and improve the availability of data. The clustered anonymous data sets are allocated different privacy budgets according to the privacy weight of the quasi-identifier attribute, and the hierarchical protection of different sensitive data is realized through the differential privacy noise mechanism, which enhances the privacy of the data. |
ICACT20230212 Question.5
Questioner: nvlinh@cs.ccu.edu.tw 2023-02-17 오전 11:54:37 |
ICACT20230212 Answer.5
Answer by Auhor 48838968@qq.com 2023-02-17 오전 11:54:37
Chrome Click!! |
| Do you compare your work with the state-of-the-art studies? |
Yes. The existing schemes also introduce clustering algorithm into data publishing privacy protection technology, but all have certain defects. The balance between privacy and availability of data has always been a hot issue in this field. The scheme in this paper reduces information loss and improves data availability by clustering at the cost of minimum information loss. At the same time, hierarchical protection of data with different sensitivity is realized through differential privacy noise mechanism, which enhances the privacy of data. |
ICACT20230212 Question.6
Questioner: nvlinh@cs.ccu.edu.tw 2023-02-17 오전 11:59:54 |
ICACT20230212 Answer.6
Answer by Auhor 48838968@qq.com 2023-02-17 오전 11:59:54
Chrome Click!! |
| What about sensitive analysis for key parameters of the algorithm |
In this paper, eight parameters K with different values are selected and divided into k-anonymous equivalence sets. It can be seen from the experiment that the sensitivity of each K value is better reflected. The BTCA algorithm proposed in this paper has smaller information loss and higher data availability with the increase of K value. |
ICACT20230212 Question.7
Questioner: m11007503@gapps.ntust.edu.tw 2023-02-19 오후 10:11:48 |
ICACT20230212 Answer.7
Answer by Auhor 48838968@qq.com 2023-02-19 오후 10:11:48
Chrome Click!! |
| Did you encounter any difficulties during the research? |
The privacy importance of each attribute in the data set is different, so it is necessary to calculate the weight of sensitive attributes in the research. However, the efficiency and usability of the algorithm in this paper may also have shortcomings when it comes to data sets containing high-dimensional sensitivity attributes. |
ICACT20230212 Question.8
Questioner: M11007513@gapps.ntust.edu.tw 2023-02-19 오후 10:06:15 |
ICACT20230212 Answer.8
Answer by Auhor 48838968@qq.com 2023-02-19 오후 10:06:15
Chrome Click!! |
| Is there more progress that can be made in this study? |
In order to make more progress in privacy protection research, the biggest challenge is to balance data privacy and availability, adjust data privacy methods based on the trade-off between utility and privacy, and create indicators to evaluate and measure the quality of model information loss and privacy protection, which is a direction of future research. |
ICACT20230212 Question.9
Questioner: m11007503@gapps.ntust.edu.tw 2023-02-19 오후 9:57:02 |
ICACT20230212 Answer.9
Answer by Auhor 48838968@qq.com 2023-02-19 오후 9:57:02
Chrome Click!! |
| Is there anything that can be done to improve this study in the future? |
The algorithm proposed in this paper can effectively reduce the information loss and insufficient privacy protection in the process of generating anonymous data sets. However, in this paper, only a single sensitive attribute in the dataset is selected for the experiment. When the dataset contains high-dimensional sensitive attributes, the algorithm efficiency and usability may have shortcomings, which is also the place to be improved in the future. |
