Publication Date: 2022/09/04
Abstract: In recent time, social media have been affected many undesirable threats. Social media provided us an open platform to connect and share our life events with others. Social media also attracted the attentions of the spammers. Spam in social media relates to undesirable, malicious and spontaneous content, shown in different ways including malicious links, massages, fake friends and microblogs, etc. With the expanding of social networks such as Instagram, Facebook, MySpace, Twitter, and Sina Weibo, etc. spammers on them are getting increasingly rampant. Social spammers consistently make a mass of phony records to misdirect the users and lead them to malicious websites and illegal content. This research is highlight features for perceiving spammers on Facebook with the help of different classifiers. Also compare the performance of different Machine Learning Algorithms (MLA) like Support Vector Machine (SVM), Multilayer Perceptron (MLP), K Nearest Neighbor (KNN) and Random Forest (RF) on machine learning tools WEKA and Rapid Miner. We use the primary data collection technique to collect the user profile data of Facebook. Lebel the data “Spam” and “Not Spam” on the basis of Engagement Rate (ER), Duplication Profile Picture and Not Human Name. The outcomes of Support Vector Machine (SVM) from the experiments is better than other algorithms on both Machine Learning Tools (MLT) WEKA and RapidMiner. The results of all algorithms are better using WEKA as compare to RapidMiner. The results will be valuable for researchers who are eager to build machine learning models to recognize spamming exercises on social media networks. Keywords:- malicious content, spammers, machine learning algorithms, Multilayer Perceptron, Random Forest, K-Nearest Neighbor, Support Vector Machine, RapidMiner, WEKA.
Keywords: No Keywords Available
DOI: https://doi.org/10.5281/zenodo.7047236
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT22AUG538.pdf
REFERENCES