A K-Anonymity Method Based on SEM (Search Engine Marketing) Price of Personal Information
Hidenobu Oguri, Noboru Sonehara · 2013
Privacy is a major concern in management of Big Data, especially for datasets that contain sensitive personal information. A model that is widely used to protect privacy is k-anonymity, which can be generally defined as a clustering method in which any record in a dataset is indistinguishable from at least (k-1) other records in the same dataset. Most of the approaches to k-anonymity suffer from huge information loss by abstraction of continuous numerical and categorical attributes that have a hierarchical structure. It is difficult to use conventional k-anonymity in actual internet services because of computational complexity and value loss from loss of information. This paper presents a k-anonymity method based on the SEM (Search Engine Marketing) price of personal information for practical use in Big Data management services. We evaluate k-anonymized qualitative data with SEM price, which is a quantitative indicator. This approach makes it possible to calculate only the necessary data and keep a k-anonymized level. Application the method to actual data shows that there is a point at which a high value can be achieved for both k-anonymity and SEM price. Developing this method will enable efficient storage of personal data and the application of k-anonymity to actual Internet services.