A Survey on Link prediction

Key Points

• The aim is to review link prediction techniques and their evolution across various domains.
• Survey of classical heuristic approaches, machine learning, and deep learning techniques
• Discussion of the impact of network domain and structure on method selection
• Evaluation of common pitfalls in reproducibility
• Identified key algorithms and methods in link prediction
• Emphasized the importance of domain-aware method selection
• Highlighted evaluation challenges affecting reproducibility