An application of big data in social media anomaly detection using weight based technique to compare performance of PIG and HIVE

Abstract

Social media is become one of the most popular application. Commonly social media is used for communication and social activities. Thus a significant amount of data is produced in these platforms and handling of these data requires advance data handling techniques thus big data is used to deal with such huge data. On the other hand, now in these days attackers and phishers are also active on social media. These attackers create fake profiles and trap the users to still their confidential and sensitive information. In this context the fake profiles are one of the serious problems in these days in social media. In this presented work a new technique for detecting the social media anomaly profile is prepared and their implementation is described in this paper. In addition of that the experimental analysis on real twitter profiles are also performed for 1200 profile features. To process these data two BIG data utilities are used namely PIG and HIVE is used. These profile features are collected from the live twitter data and evaluation of different profiles. The experimental results are compared for both the utilities (i.e. PIG and Hive) to demonstrate the successfully identification of legitimate and anomaly profiles.