Title: Correlated Feature Selection for Tweet Spam Classification using Artificial Neural Networks

Abstract: Identification of spam messages is a very challenging task for social networks due to its large size and complex nature. The purpose of this paper is to undertake the analysis of spamming on Twitter. To classify spams efficiently it is necessary to first understand the features of the spam tweets as well as identify attributes of the spammer. We extract both tweet based features and user based features for our analysis and observe the correlation between these features. This step is necessary as we can reduce the training time if we combine the features that are highly correlated. To perform our analysis we use artificial neural networks and train the model to classify the tweets as spam or non-spam. Using Correlational Artificial Neural Network gives us the highest accuracy of 97.57\% when compared with four other classifiers SVM, Kernel SVM, K Nearest Neighbours and Artificial Neural Network.