Generalized Motif-based Naive Bayes for Sign Prediction
Research Paper#Network Science, Machine Learning, Sign Prediction🔬 Research|Analyzed: Jan 3, 2026 19:37•
Published: Dec 28, 2025 03:53
•1 min read
•ArXivAnalysis
This paper addresses the limitations of traditional motif-based Naive Bayes models in signed network sign prediction by incorporating node heterogeneity. The proposed framework, especially the Feature-driven Generalized Motif-based Naive Bayes (FGMNB) model, demonstrates superior performance compared to state-of-the-art embedding-based baselines. The focus on local structural patterns and the identification of dataset-specific predictive motifs are key contributions.
Key Takeaways
- •Proposes a generalized framework for sign prediction in signed networks.
- •Models node heterogeneity to improve performance.
- •FGMNB model outperforms state-of-the-art baselines.
- •Highlights the importance of local structural patterns and dataset-specific motifs.
Reference / Citation
View Original"FGMNB consistently outperforms five state-of-the-art embedding-based baselines on three of these networks."