Preventing A False Light Caused by k-anonymity with Mathematical Modling and Optimization

2015 Fiscal Year Final Research Report

• Project/Area Number: 26540041
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Multimedia database
• Research Institution: The University of Tokyo
• Principal Investigator: NAKAGAWA Hiroshi 東京大学, 情報基盤センター, 教授 (20134893)
• Co-Investigator(Kenkyū-buntansha): SATO Issei 東京大学, 新領域創成科学研究科, 講師 (90610155)
• Project Period (FY): 2014-04-01 – 2016-03-31
• Keywords: プライバシー / k-匿名化 / 濡れ衣 / 個人情報 / パーソナルデータ / 位置情報 / 属性情報

Outline of Final Research Achievements

In the field of privacy preserving data mining, k-anonymity is a representative model for protecting privacy. However, when people see k-anonymized data, a person who provides his/her data is misleadingly suspected as a bad guy due to the information which actually has nothing to do with him/her. We define such problem as a false light caused by k-anonymization, and define a record which has an attribute causing a false light as a sensitive record. We propose k-anonimisation algorithms which pay attention to sensitive records in order to prevent a false light. We deal with two cases: location information preservation and more general information including categorical information. In the experiments, we confirmed that proposed method can decrease a probability of occurrence of a false light.

Free Research Field

情報工学