An Intelligent Intrusion Detection Model for Prevention of Attacks on Social Media Network Platform

Abstract

Social media network platform (SMNP) connects people worldwide to interact, share content, and engage in discussions of mutual interest that know no geographical boundaries. Although the gain is incredible, most traditional crimes now have digital equivalence which threatens the effectiveness of virtual socialization. The aim of this research work was to use machine learning classifiers to develop a detection model that will prevent attacks on SMNP. The study adopted the analytical method. Scikit learn version 1.1.1 was used to evaluate the performance of three machine learning classifiers in terms of accuracy, precision, detection rate, F1-Scoreand Receiver Operating Characteristics (ROC) curve using a generated dataset of size 36,590 feature-set from twitter platform. Results obtained from the three ML classifiers in terms of accuracy, precision, detection rate, F1 Score, and ROC curve were: for KNN 80.14%, 69.35%, 73.64%, 71.42% and 85.53% respectively; for LR 80.98%, 72.12%, 74.05%, 73.07% and 86.59% respectively; and for DT 79.76%, 72.05%, 71.58%, 71.81% and 78.05% respectively. LR proves to be the most efficient classifier suitable for exposing attacks on social media network platform.