Prevention and Detection Attack of Social Media Networks Using Machine Learning Methods

preventing malicious and cyberbullying activity, which is primarily carried out by fake profiles, spam messages, and other such actions. The privacy practices of the social network communities are questioned by these actions. These phony profiles are to blame for the dissemination of incorrect information on social networks. Much study has been done in this field to identify fraudulent profiles, duplicate posts, spam accounts, and bot accounts. The majority of bogus accounts were successfully identified by employing a machine-learning system. This paper is a variety of feature selection techniques as well as a variety of machine learning-based methodologies are employed to identify fake accounts that can deceive users. Eight feature selection strategies were combined with five machine learning methods to preprocess the dataset and identify bogus accounts, The feature selection process (which uses multiple methods to rank features) includes eight feature selection processes (Chi-Square Test Feature Selection, ANOV A Feature Selection, Mutual Information Feature Selection, Logistic Regression Feature Selection, Additional Tree Feature Selection, Embossed Feature Selection, mrMR features Selection and, Light GBM Trait Selection), all outputs are evaluated by finding similar features in the process of choosing between the outputs of the eight algorithms to find the best features to be used by the classifier for best performance. A classification method consisting of five machine learning algorithms: Support Vector Machine (SVM), Gaussian Naive Bayes (GNB), Stochastic Gradient Descent (SGD), Random Forest (RF), and Decision Tree (DT) was used and tested, the best of which proved to be the best in the identification process we mentioned.