GAIN RATIO BASED FEATURE SELECTION METHOD FOR PRIVACY PRESERVATION

Abstract
Privacy-preservation is a step in data mining that tries to safeguard sensitive information from unsanctioned disclosure and hence protecting individual data records and their privacy. There are various privacy preservation techniques like k-anonymity, l-diversity and t-closeness and data perturbation. In this paper k-anonymity privacy protection technique is applied to high dimensional datasets like adult and census. since, both the data sets are high dimensional, feature subset selection method like Gain Ratio is applied and the attributes of the datasets are ranked and low ranking attributes are filtered to form new reduced data subsets. K-anonymization privacy preservation technique is then applied on reduced datasets. The accuracy of the privacy preserved reduced datasets and the original datasets are compared for their accuracy on the two functionalities of data mining namely classification and clustering using naïve Bayesian and k-means algorithm respectively. Experimental results show that classification and clustering accuracy are comparatively the same for reduced k-anonym zed datasets and the original data sets.

Authors
R.Praveena Priyadarsini1, M.L.Valarmathi2 , S.Sivakumari3
Avinashilingam Deemed University for Women, Tamil Nadu, India1 , Government College of Technology, Tamil Nadu, India 2 , Avinashilingam Deemed University for Women, Tamil Nadu, India3

Keywords
Privacy Preservation, Data Mining-Anonymity, Feature Subset Selection, Gain Ratio
Published By :
ICTACT
Published In :
ICTACT Journal on Soft Computing
( Volume: 1 , Issue: 4 )
Date of Publication :
April 2011

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.