An efficient approach for spammer detection on Twitter and their behavior analysis
K Anaswara, A. Saleema, V. Thanammal Indu · 2022 IEEE International Conference on Signal Processing, Informatics, Communication and Energy Systems (SPICES) · 2022
Twitter is the 4th most popular social networking site and almost 353 million are using Twitter worldwide. The main problems faced by Twitter users are spam, malicious automation, and platform manipulation. Twitter’s developer API is exposed to make it easy to interact. So, the anti-spam system on Twitter is known by spammers. Here, we are using some features that are related to account details of Twitter users with two different data sets having one lakh data for each. We propose a method to classify spam and non-spam users using five machine learning classifiers. We evaluated the performance using accuracy, precision, and recall. Then we are proposing a fuzzy-based approach for user behavior analysis, inspired by the Five-factor model of behavior theory. Only a few works are discussing the behavior of the spammer, and we found that there exists a relationship between the traits of a person and spamming.