What SAS procedures can be implemented for L-diversity or T-closeness

[ShelleySessoms]

It's best to provide this community with as much information as possible, in order to get a helpful reply. Here are a few tips to getting a fast, helpful reply:

How to ask a good question and receive a fast reply

Experts who respond on the communities have tremendous experience and intuition, but they aren't mind readers. You have to form good questions if you want to receive a helpful response. Here are some tips for success:

• Use a precise subject line. Try to include your goal, error message, SAS procedure name, function -- whatever keywords will help an expert to "pick up" your question as something he/she could answer. (Pro tip: "Urgent help needed!" or "SAS question" are not effective subject lines.)
• Share example data. Many questions can be answered properly only when the responders can see the "shape" and characteristics of your data. Don't share anything proprietary, of course.
• Show what you have tried. Community members love to nudge you towards the proper solution, but it helps if you share what you've already tried and if you hit any walls...explain. If you have special constraints (must use a older version of SAS, certain product set, etc), share that too.
• Search on communities site first, before posting a new question. The act of entering a new question helps with this because you'll see the subject line "autocomplete" with suggested matching topics, even before you post. That's another reason that the first tip (precise subject line) is so important.
• Post into most appropriate board. There are boards for most SAS products, and these are monitored regularly by experts who specialize. Posting on the correct board helps your topic to be seen by the best experts.
• When you receive a helpful reply, come back and mark it as an Accepted Solution, or at least click Like for the replies that are helpful. This action will help you and others to find the answer in the future.

[mssh2712]

Sorry, I’ll change they I write to post a question. But this question is been posted while ago in Sas statistical procedure. No reply. Thanks

[ballardw]

---

@mssh2712 wrote:
Sorry, I’ll change they I write to post a question. But this question is been posted while ago in Sas statistical procedure. No reply. Thanks

---

Perhaps you need to define L-diversity and T-closeness and or post a link to what and how these are used. There are many groups that have internal jargon that call something another groups uses all the time by a different name. I used to get driven nuts when tutoring business students where they would bring in three or four "different" business terms but all came down to finding a local minimum or maximum but they changed the name depending on what the called the base curve. The methods for finding a local min/max for a polynomial equation are the same regardless of whether the equation models "price" "demand" "yield" or what have you.

I suspect that may be your case.

If you can show us what an L-diversity looks like some one may recognize that as result X when using Option Y in Proc Z.

Best would be a worked example to reference.

[Reeza]

As far as I know there is no specific procedure for L-diversity or T-closeness that deal with this directly. 

You could either use PROC IML to roll your own, or depending on how you define the calculations use PROC FREQ and some BASE SAS coding to get you there. 

This isn't a trivial problem and there are entire software applications built to help mask/anonymize data. 

If you have specific further questions please feel free to post more details.