| Summary: | 碩士 === 國立屏東科技大學 === 資訊管理系所 === 100 === Privacy preserving is becoming a crucial issue with the development progress of data mining techniques. An important challenge in data mining is to enable the legitimate usage and sharing of data mined information while at the same time guaranteeing proper protection, especially, from the inferences of the original sensitive data. k-anonymity is a property that models the protection of released data against possible re-identification of the respondents to which the data refer. In this paper, we propose an effective k-anonymity solution for classification that is a fundamental problem in data analysis. The proposed scheme conducts SVM and a revised C4.5 along with k-anonymity.
The key features of our scheme include the betterment of data quality on information loss, lightweight overhead in computation, and the accomplishment of privacy-related goals. Finally, we will give experiments to show how the proposed solution works.
|
|---|
