Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/125427
Title: Enhancing Privacy in Online Social Networks using Data Analysis
Researcher: Srivastava Agrima
Guide(s): Geethakumari G
Keywords: Social Networking Sites, Data Analysis, Security
University: Birla Institute of Technology and Science
Completed Date: 18/08/2015
Abstract: Online Social Networks (OSNs) are widely used social computing platforms where on a newlinedaily basis huge quantum of personal information is shared by the users. This personal information newlineis sensitive in nature and could be misused by the adversaries resulting in privacy newlineviolations. Hence, we need to study and enhance the present privacy mechanisms in order newlineto reduce the chances of unwanted information disclosures in an OSN. In this research newlinework, we have explored data privacy from the perspectives of OSN users, their online connections newlineand the service providers. We identified the loopholes in the existing mechanisms newlineand proposed efficient data privacy enhancing algorithms. The proposed measures would newlinehelp the users understand the privacy risks and would enable them to prudently share their newlinedata online. newlineThe three main challenges that are being addressed in the thesis are measuring OSN newlineusers data privacy, enabling selective sharing of sensitive OSN data and preventing sensitive newlineOSN data from inference attacks. We measured users data privacy using the theory newlineof psychometrics. We proposed a privacy settings recommender system to recommend newlineappropriate privacy settings for the OSN users profile. We analyzed the effect of node s newlinetopology on sensitive information spread and proposed the trust enhancing model to refine newlinethe existing trusted community. An efficient partial edge set removal algorithm is proposed newlineto reduce the accuracy with which the attacker could infer the sensitive attributes. We also newlineproposed a privacy utility trade-off algorithm that could offer maximum utility and minimum newlineprivacy loss in the released anonymized datasets. Our work aims to build efficient newlinedata privacy enhancing solutions which could protect the users data against privacy attacks newlineand make OSNs a privacy preserving platform for data sharing.
Pagination: 
URI: http://hdl.handle.net/10603/125427
Appears in Departments:Computer Science & Information Systems

Files in This Item:
File Description SizeFormat 
synopsis.pdfAttached File972.37 kBAdobe PDFView/Open
thesis_final.pdf2.1 MBAdobe PDFView/Open


Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.