User Identification Across Multiple Social Networks Based on Naive Bayes Model

Area of Science:

• Computer Science
• Network Science
• Data Science

Background:

• User identification across multiple social networks (UIAMSNs) is crucial for many applications.
• Existing feature-based methods often lack a strong theoretical foundation for matching degrees.

Purpose of the Study:

• To propose a theoretically grounded algorithm for UIAMSNs using a naive Bayes model.
• To enhance the accuracy and efficiency of cross-platform user identification.

Main Methods:

• Developed a naive Bayes model-based matching degree index to quantify contributions of matched node pairs (MNPs).
• Formulated the matching degrees of unmatched node pairs (UMNPs) using matrix products for efficiency.
• Employed a recursive process for iterative prediction of UMNPs with limited prior information.

Main Results:

• The proposed User Identification based on Naive Bayes Model (UI-NBM) accurately measures contributions of MNPs.
• Matrix formulation significantly reduces time complexity and provides a compact expression.
• Recursive approach enables effective prediction of UMNPs even with sparse initial data.

Conclusions:

• The UI-NBM offers a theoretically sound and efficient approach to UIAMSNs.
• The method demonstrates superior performance compared to baseline approaches on synthetic and real-world data.
• This work provides a robust framework for cross-platform user identification.