Comprehensive Survey on Big Data Privacy Protection

Mohammed Binjubeir,Ali Safaa Sadiq,Abdulghani Ali Ahmed,Muhammad Khurram Khan,Mohd Arfian Bin Ismail

doi:10.1109/access.2019.2962368

Abstract

In recent years, the ever-mounting problem of Internet phishing has been threatening the secure propagation of sensitive data over the web, thereby resulting in either outright decline of data distribution or inaccurate data distribution from several data providers. Therefore, user privacy has evolved into a critical issue in various data mining operations. User privacy has turned out to be a foremost criterion for allowing the transfer of confidential information. The intense surge in storing the personal data of customers (i.e., big data) has resulted in a new research area, which is referred to as privacy-preserving data mining (PPDM). A key issue of PPDM is how to manipulate data using a specific approach to enable the development of a good data mining model on modified data, thereby meeting a specified privacy need with minimum loss of information for the intended data analysis task. The current review study aims to utilize the tasks of data mining operations without risking the security of individuals’ sensitive information, particularly at the record level. To this end, PPDM techniques are reviewed and classified using various approaches for data modification. Furthermore, a critical comparative analysis is performed for the advantages and drawbacks of PPDM techniques. This review study also elaborates on the existing challenges and unresolved issues in PPDM.

Highlights

Various organizations in different sectors have been striving to make their data electronically available
Data is generated from several sources, such as business sales records, sensors used in the internet of things, social media, medical patient records in healthcare organizations, video and image archives [4]
The existing privacy-preserving data mining (PPDM) techniques and are intensively reviewed and classified based upon their methods that used data modification approaches, which represented the main contribution of this study that will help researchers in this field having comprehensive understanding of PPDM

Summary

INTRODUCTION

Various organizations in different sectors (e.g., government, banking, medical, and insurance sectors, as well as public and private institutions) have been striving to make their data electronically available. The current study adopts the definition of privacy in the contexts of content and interaction [12], [18], which is related to research path in terms of the collection and analysis of individual data This can be valuable in boosting the effectiveness of organizations or support prospective plans. Numerous endeavors have been dedicated to privacy, which involve the preservation of individuals’ information using data mining algorithms, to avert the disclosure of individuals’ identities or sensitive data in the course of knowledge discovery [21]. This paradigm is referred to as PPDM. Given the data modification is the main focus of this review paper, it is further reviewed and deeply analyzed

DATA MODIFICATION APPROACHES

PRIMARY TASKS OF DATA MINING

Findings

CONCLUSION