Extreme Learning Machine for Spammer Detection and Fake User Identification from Twitter

The proliferation of online social networking sites (OSNs) platforms like Twitter has led to an increase in spam and the presence of fake user accounts, posing significant challenges to user experience and platform resources. This research focuses on the detection of spammers on Twitter as well as the identification of fraudulent users. The research begins by pre-processing the Twitter dataset to remove noise and irrelevant information. Next, Multinomial Naive Bayes (MNB) is employed to extract informative features from the preprocessed dataset, considering various textual, linguistic, and user-specific characteristics. After the characteristics have been collected, they are utilized to train a classifier called an Extreme Learning Machine (ELM), which is able to efficiently differentiate between spam accounts and real user accounts. Extensive tests are run on an actual Twitter dataset to assess the effectiveness of the suggested methodology. The results demonstrate that the combination of MNB feature extraction and ELM classification achieves high accuracy, precision, recall, and F1-score in identifying spammers and fake users on Twitter.