Indian Journal of Science and Technology

Abstract

Objectives: In recent scenario, online social networks such as Face book, Twitter and Google+ have become one of the fastest emerging e-services. There are several issues affected these e-services. Since it is emerging service and reliability to communicate, in social networks privacy is often a key concern by the users. Since millions of people are willing to interact with others, it is also a new attack ground for malware creators. Some users and pages spreading malicious content and sending spam messages by taking advantage on the users’ inherent trust in their relationship network. Methods:This proposed work handles the most prevalent issues and threats targeting different social networks recently. And finally finds the authentication scheme for those attacks. This proposes a detecting and blocking scheme for social sites using data mining techniques. Findings: This system helps to detect suspicious URLs for social network by considering the following parameters, i).Text and keywords appears in the URL. ii). URL descriptions iii). Detection of scam messages which is done in manual script attacks on social sites. Application/Improvement: This performs two techniques which are message filtering and MLE (Maximum Likelihood Estimation).
Keywords: Feature Selection, Machine Learning, Security, Social Network