A Novel Quantum Beta Distributed Multi-Objective Particle Swarm Optimization Algorithm for Fake Accounts Detection
Description
The proposed Quantum Beta Distributed Multi-Objective Particle Swarm Optimization (QB-MOPSO) algorithm is designed to: Perform feature selection by minimizing both feature dimensionality and classification error. Detect fake accounts on social networks using machine learning classifiers such as Random Forest, Support Vector Machine (SVM), Naïve Bayes, and Neural Networks. Combine quantum-behaved exploration with beta-distributed exploitation to enhance convergence and detection accuracy. Dataset Description: We use two publicly available Twitter datasets collected by Cresci et al. (2017), which are widely adopted in fake account detection research. These datasets consist of labeled Twitter accounts, categorized as either genuine users or social spambots. The published datasets were normalized and used in our experimental study. Dataset 1 contains 1,982 accounts and 4,061,598 tweets, including genuine users and Social Spam Bot 1 (focused on Italian political retweets). Dataset 2 includes 928 accounts and 2,628,181 tweets, containing genuine users and Social Spam Bot 3 (targeting Amazon product promotion).