L-Diversity explained

[@tonyaradzwachivandire1608]

Thank you! I was failing to understand what "well represented values" meant concretely. This video cleared up the confusion!

[@andybarbarian]

My prof gave an entire rambly 2h lecture on this topic and left me thoroughly confused. You explained it clearly in less than 4 minutes. Thank you so much! Subscribed and will now continue the playlist :D Btw, the database anonymization playlist has the fist two videos switched, this one should come after k-anonymity as you point out at the beginning ^^

[@SecPrivAca]

Thanks for the heads up! And glad you enjoyed the video.

[@AndriiUshakov]

Hey, man, by any chance you do not study at the uni.lu? :)

[@andybarbarian]

@@AndriiUshakov nah, unibe though ^^ just had my exam this week and went great thanks to this guy! you?

[@SecPrivAca]

@@AndriiUshakov No, I'm in Germany. But you probably meant Andy ;)

[@MooseOnEarth]

0:49 - The "red frame" for the quasi identifier should only cover the first two attributes (Age and ZIP). The third one (Disease) is the *sensitive* attribute and is not part of k-anonymity. This is an important distinction. In your videos attribute columns and columns with sensitive information could be colored differently in order to make this distinction.

[@moazzammehmood9643]

Isn't l-diversity used to overcome skewness attacks on p-sensitive, k-anonymity?

[@SecPrivAca]

When L-diversity was introduced, it was directly addressed to overcome distribution issues in k-anonymous databases. The paper is available and relatively readable, if you're interested to know more: dl.acm.org/doi/pdf/10.1145/1217299.1217302?casa_token=T2ib53ZLitIAAAAA:Ftcww6ad8dz59oJmV7nOWFNLYTdgyorY2EGBJDFqylcnFJZszVygF1HxgPqdZwqVfY1yiMVrSlLNKw