11 lines
1.4 KiB
TeX
11 lines
1.4 KiB
TeX
\section{Summary}
|
|
\label{sec:lmdk-sum}
|
|
In this chapter, we presented \emph{{\thething} privacy} for privacy-preserving time series publishing, which allows for the protection of significant events while improving the utility of the final result compared to user-level differential privacy.
|
|
We proposed three schemes for {\thething} privacy, and quantified the privacy loss under temporal correlation.
|
|
Furthermore, we designed a module to enhance our privacy notion by protecting the actual timestamps of the {\thethings}.
|
|
We differ the experimental evaluation of our methodology to Chapter~\ref{ch:eval} we experiment with real and synthetic data sets to demonstrate the applicability of the {\thething} privacy models by themselves (Section~\ref{sec:eval-lmdk-sel}) and in combination with the {\thething} selection component (Section~\ref{sec:eval-lmdk}).
|
|
|
|
%Our experiments on real and synthetic data sets validate our proposal.
|
|
%In the future, we aim to investigate privacy-preserving {\thething} selection and propose a mechanism based on user-preferences and semantics.
|
|
% \kat{Advertise your work! Say what is cool about the work and how it differs from the others! Mention also the summary for selection of events. The discussion for the experiments and future work you postpone for the respective sections, you may though make reference to specific experiments to support your claims. }
|