Data Security: Modified Privacy-Preserving Data Mining Algorithm

• Main Author: Leshchenko Illia
• Other Authors: Shi-Jinn Horng
• Format: Others
• Language: en_US
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/60128231051349181874

碩士 === 國立臺灣科技大學 === 資訊工程系 === 104 === Nowadays the majority of people in developed countries are using the Internet. Therefore, all of them give their personal data to third-parties, which can use it on specified conditions. However, none of the Internet websites are completely protected from malicious users, especially when those third-parties are using data mining technique, which is pretty common now. This thesis focuses on inventing a modified algorithm to provide better personal data security comparing to existing ones. This algorithm reduces a leakage of personal information for public use. Modified Privacy-Preserving Data Mining (MPPDM) algorithm works as follows: when data owner wants to perform data mining and publish personal information about customers, he must provide personal data anonymity first to avoid disclosure of user identity. For this purpose, he uses MPPDM and afterward can post the result for public use. The result of the algorithm is quite good and looking better than results of existing algorithms. Specification of MPPDM provides a chance to change the level of anonymity manually if it is needed. Compared to the basic PPDM, the proposed MPPDM shows better advantages.