Artificial Intelligence Enhanced Hybrid Fuzzy Technique in Social Media Mining for Anomaly Detection

Abstract

The investigation discusses Horizontal Anomalies in social media, emphasizing their multidimensionality and associated hazards. Traditional anomaly detection methods may fail to detect Horizontal Anomalies, demanding more advanced algorithms. Fuzzy logic, which can simulate uncertainty, is used alongside the K-means clustering technique to partition data. The anomaly identification process uses high membership numbers to determine activity level and engagement rate. Line plots illustrate membership values, whereas cluster centroids indicate the clusters. This multidisciplinary method, which includes fuzzy logic, clustering, anomaly detection, and visualization approaches, manages the complexities of Horizontal Anomaly detection, hence improving social media integrity and user security. Flawless identification of horizontal abnormalities in social media mining is crucial for maintaining platform integrity, safeguarding user privacy and security, and countering fraudulent activity. Despite its relevance, horizontal anomaly detection is one of the least researched aspects of social media mining. This study proposes a fuzzy logic-based technique supplemented with K-means clustering to detect horizontal abnormalities accurately and efficiently. The goal is to create a robust system capable of detecting unusual user activity across various dimensions, helping to progress anomaly detection techniques in social media mining.