Splošna definicija diferencirane zasebnosti

delo diplomskega seminarja

Metod Jazbec (Author), Aljoša Peperko (Mentor)

Abstract

V delu predstavimo diferencirano zasebnost. Gre za matematično definicijo zasebnosti pri javni objavi ter rudarjenju podatkov. Predstavljena je splošna definicija v kontekstu metričnih prostorov in verjetnostne mere, ki omogoča enotno obravnavo različnih vrst podatkov. Pokažemo nekaj osnovnih izrekov, ki omilijo zahteve definicije. Obravnavan je Laplaceov mehanizem za numerične podatke. Podana je izpeljava spodnjih mej za največjo napako zasebnih odzivnih mehanizmov. V nadaljevanju se osredotočimo na funkcijske podatke. S pomočjo teorije Gaussovih procesov in Hilbertovih prostorov z reprodukcijskim jedrom pokažemo uporabo diferencirane zasebnosti na primeru jedrne cenilke gostote. Osnovne mehanizme implementiramo in predstavimo rezultate.

Keywords

matematika;diferencirana zasebnost;odzivni mehanizem;metrični prostori;funkcijski podatki;verjetnostna mera;

Data

Language:	Slovenian
Year of publishing:	2018
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FS - Faculty of Mechanical Engineering
Publisher:	[M. Jazbec]
UDC:	519.8
COBISS:	18418009
Views:	855
Downloads:	283
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	General definition of differential privacy
Secondary abstract:	We introduce the concept of differential privacy, mathematical definition for privacy preserving data publishing and data mining. General definition in context of metric spaces and probability measure is given. Further, we present some theorems which help to alleviate the requirements of described definition. Laplace mechanism for numerical data and lower bounds on errors of response mechanisms are presented. We later turn focus to functional data. Using Gaussian processes and Reproducing Kernel Hilbert Spaces we present how differential privacy is used for privatization of density kernel estimator. Most of the described mechanisms are also implemented and results are presented at the end
Secondary keywords:	mathematics;differential privacy;response mechanism;metric spaces;functional data;probability measure;
Type (COBISS):	Final seminar paper
Study programme:	0
Thesis comment:	Univ. v Ljubljani, Fak. za matematiko in fiziko, Oddelek za matematiko, Finančna matematika - 1. stopnja
Pages:	31 str.
ID:	10956109