A personalized privacy-preserving method for multi-view clustering mining

Abstract

The invention discloses a personalized privacy protection method for multi-view cluster mining, which belongs to the technical field of information security. The present invention proposes a privacy partial order topology classification algorithm (PT, Privacy Topology). Aiming at the sensitivity difference representation problem of different privacy data, first define the privacy relationship and construct a privacy partial order set, and then design a topology classification algorithm for privacy data to solve the privacy line Prologue. For multiple views of privacy data, multi-view clustering is performed on views such as raw data, privacy degree, tuple sensitivity, and privacy line sequence set. A clustering-oriented personalized anonymous algorithm (PPOC, Personal Privacy Oriented Classtering) is proposed, through the variable k-anonymous strategy, using the privacy protection algorithm oriented to multi-view clustering to meet the individual needs, to realize the classification of different clusters and the same Different tuples within the cluster impose different degrees of personalized protection operations.

Description

technical field [0001] The invention relates to a personalized privacy protection method for multi-view cluster mining, and relates to the technical field of information security. Background technique [0002] With the development of digital technologies such as the Internet, the Internet of Things, and smart cities, various data collection devices such as sensors and mobile terminals store various information about human clothing, food, housing, and transportation in digital form, thus giving birth to the era of big data. For the first time, data, as a resource, has received great attention from social entities such as governments, enterprises, and academia. [0003] However, in the process of data use, personal privacy information may be leaked. In the process of discovering the potential value of data, how to protect the privacy of individuals, especially how to avoid privacy leakage caused by data mining, is a key issue that data science needs to solve urgently. Privat...

AI Technical Summary

Problems solved by technology

[0007] (1) Existing algorithms are almost designed for the needs of data release, and have not been involved in privacy and security issues that may be caused by cluster mining;

[0008] (2) The existing personalized privacy protection algorithm does not comprehensively consider the differences in the privacy awareness of users and the importance of different attributes, the degree of personalization needs to be improved, and the information loss is relatively large

Images