Location Data Anonymization Retaining Data Mining Utilization.

Location information, such as customers' home addresses, is essential for data mining tasks. On the other hand, it is sensitive private information. In most countries, when requesting a third party for data mining, it is legally required to anonymize personal information such as home addresses in the database. However, conventional anonymization methods significantly lose the usefulness that location information has inherently. In this paper, we proposed an anonymization method of location retaining important locational features. In the proposed method, each address is replaced with a ranking value of the distance from each facility that is important in terms of location, such as a station or a supermarket. We examined our method and confirmed that the important rules mined from non-anonymized data could also be mined from our anonymized data.