A multi-classifier approach for twitter spam detection using innovative ann-fdt algorithm

Abstract

Nowadays, various social media platforms are available in Internet like Facebook, Twitter and Instagram for uniting the people. Twitter is one among the most famous platform in social media due to its available information among users. Users allows to find new friends and update their latest information and activities. Twitter is using Google Safe-browsing to detect the spam URL and block spam links. Due to the presence of advanced API which enables to read and write the data in Twitter, different kinds of spammers are attracted in the Twitter. There are various existing researches applied various machine learning techniques to determine the twitter spam. However, there is no comprehensive evaluation on their algorithms and lack of accuracy in large dataset. To rectify these issues, this research proposed hybrid method with the combination of Artificial Neural Networks with Fuzzy Decision Tree (ANN-FDT). The proposed classifier classified the span and non-span tweets based on the labels. For experimental analysis, the proposed classifier applied on large dataset of 600 million public tweets. The performance of proposed algorithm is evaluated by means of measures like accuracy, TPR, FPR and F-measure. From the results it can be seen that the proposed technique has improved performance.