An Intelligent Centrality Measures for Influential Node Detection in COVID-19 Environment.

Area of Science:

• Social Network Analysis
• Machine Learning
• Information Security

Background:

• Spamming on social networks poses significant user challenges.
• Identifying influential users who spread spam is a critical research problem.
• Spam impacts users socially and psychologically.

Purpose of the Study:

• To propose a novel method for detecting influential spam-spreading users in online communities.
• To enhance the identification of spam accounts versus genuine ones.

Main Methods:

• Community detection using Laplacian Transition Matrix and hashtag analysis.
• Identification of influential nodes via intelligent centrality measures.
• Classification of user influence intensity using machine learning algorithms (SVM, PCA).

Main Results:

• The proposed method achieved 98.6% accuracy using Support Vector Machines (SVM) and Principal Component Analysis (PCA).
• New centrality measures and scores like Normalized Mutual Information (NMI) and Root Mean Square (RMS) were evaluated.
• SVM and PCA outperformed linear regression in classifying user influence.

Conclusions:

• The developed approach effectively identifies influential users spreading spam.
• This research aids in distinguishing between spammy and genuine social media accounts.
• The findings contribute to mitigating the negative impacts of spam in online communities.