Efficient Spam Detection on X (formerly Twitter) : A Hybrid Artificial Neural Network and Fuzzy Decision Tree Approach

Abstract

These days, there are a plethora of online social media sites that bring people together, such as Instagram, X (Twitter), and Facebook. The abundance of user-generated content on X Platform has made it a leading social media platform. Users are able to connect with one another, share what they're up to, and discover new pals. Twitter detects spam URLs and blocks them using Google Safe-browsing. X (Twitter) attracts several types of spammers since it has a sophisticated API that allows users to read and publish data. Many previous studies have used different machine learning algorithms to identify spam on X. On the other hand, their methods have not been well tested, and they have shown to be inaccurate when applied to huge datasets. A hybrid approach is created by integrating Artificial Neural Networks along with Fuzzy Decision Trees was proposed as a solution to these problems in this study. According to the labels, the suggested classifier distinguished between span and non-span tweets. A massive dataset consisting of 600 million public tweets was used to test the suggested classifier. To assess the new calculation's presentation, metrics such as accuracy, F-measure, TPR and FPR are employed. The results shows that the our novel strategy is more reliable and powerful.