Thesis Details
Anonymizace dat
The goal of this work is to evaluate reasons and cases of usage of anonymization in context of privacy protection in data mining. Privacy Preserving Data Mining (PPDM) has been an area of research since 1991. It has also been discussed at many international conferences. Currently the research is mainly directed towards development of technical methods, such as application of cryptography or the development of specialized algorithms to meet security and privacy requirements for different data mining methods, such as classification or categorization. So far PPDM has found application in only a few cases. For example, one of them can be medical research to protect patient's privacy. In all cases when data mining is applied to personal data, this process has to be in compliance with data protection legislation. This results in responsibilities for data controllers, technical operators and others involved in those business or governmental processes where data mining plays a role. In connection with this problem we meet with k-anonymization methods, especially with class of these methods such as clustering-based. These methods are able to achieve high quality anonymization and thus have a great application potential.
Data mining, anonymization, privacy protection, k-anonymization, clustering