期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2013
卷号：58
期号：3
出版社：Journal of Theoretical and Applied
摘要：In data mining applications, privacy plays an imperative role. This has triggered the development of many privacy preserving data mining techniques. To facilitate privacy preservation in data mining or machine learning algorithms over horizontally partitioned or vertically partitioned data, many protocols have been proposed using SMC and various secure building blocks. Our previous works focused on preserving privacy by adapting individually adaptable perturbation model, which enables the individuals to choose their own privacy levels. But the downside is that it does not discover the computational results for privacy properly. This paper proposed a feature selection with privacy preservation in multi-partitioned dataset. Data can be sealed for privacy by perturbation technique as pseudonym name. In multi-partitioned data evaluation, it creates classification of data and selection of feature for data mining decision model which construct the structural information of model in this paper. The purpose of gain ratio method has taken in this paper to enhance the privacy in multi-partitioned data set. All features don�t require protecting the confidential data for best model. The data representation for privacy preserving data mining has taken to increase the data mining technique to construct finest model without breaking the privacy individuals. An experimental evaluation is conducted to estimate the performance of the proposed enhanced privacy preservation with perturbed data using feature selection [EPPDFS] in multi-partitioned datasets demonstrated by diverse experiments conducted on both synthetic and real-world data sets.
关键词：Privacy Preserving; Data Mining; Perturbed Data; Feature Selection
