Abstrato

Enhanced Slicing Technique for Improving Accuracy in Crowdsourcing Database

T.Malathi, S. Nandagopal

In recent years, privacy preserving has seen rapid growth which leads to an increase in the capability to store and retrieve personal dataset without revealing sensitive information about the individuals. Different techniques have been proposed to improve accuracy in crowdsourcing database. Anonymization techniques such as, generalization and bucketization, are designed for improving accuracy in privacy preserving method. But the malicious workers can hack the private information of the user and misuse it. Recent work has been shown that k-anonymity for generalization losses considerable amount of information especially for higher dimensionality data. l-diversity for bucketization does not able to prevent membership disclosure. In this paper we introduce a novel technique called overlapped slicing, which partitions the data in both horizontal and vertical manner. Slicing preserves better data utility than generalization and bucketization techniques. As an extension we proposed a technique called overlapped slicing, in which an attribute is divided into more than one column. The release in each column consists of more attribute correlations. Important advantage of this work is to handle high-dimensional data and also preserves better privacy than the previous techniques.