TWITTER ACCOUNT PREDICTION USING MACHINE LEARNING

Today's society is facing numerous cyber security related problems. In advancement of machine learning, C4.5 decision tree algorithm is applied in existing system to predict fake and clone profile in social media. It addresses cyber security needs, but it predicts clone profiles only using blank profile information. So, it achieves only 50-60 percent of accuracy, since most of the fake profiles cannot be detected using this system. Hence system has to be proposed for predicting clone profiles efficiently. This task proposes Random Forest algorithm along with Decision tree algorithm, which efficiently finds fake profile using rule based, attribute-based similarity feature, network-based feature. The system is aimed to achieve more than 80% of accuracy. Cloned profiles can also be detected using their comments and activities performed in social media. This is predicted using RF, and NN algorithm.