In the context of location privacy, since k-anonymity is often used as the priva
ID: 3720573 • Letter: I
Question
In the context of location privacy, since k-anonymity is often used as the privacy protection model, can clustering be used to provide the MBR? Are there any challenges unique to the location privacy context for this?
Consider the following method for achieving location privacy: For each person, you find the k-closest people around that person and then you generalize the location to be the MBR over the k people around you. Is this data dependent cloaking or space dependent cloaking. If space dependent cloaking, is it fixed-grid or adaptive-grid? Why?
Are all data-driven cloaking methods susceptible to location distribution attack? Why or why not?
Explanation / Answer
Answer is as follows:
According to Chegg Guidelines , we are allowed to solve only first part of the question, so please repost other two parts...
MBR i.e.Minimum Bounding Rectangle is the smallest region which covers all the events under the cloak where region include the maximum and minimum values of events.
We use clustering , the data is divide into different groups where the object in each cluster is similar to the objject of other cluster. We can use K-means clustering method to make the group of nearest messages sent by users according to their locations.
So we can use clustering to provide the MBR.
There are some challenges like :
if there is any query please ask in comments...