Link Prediction in Social Networks

This work presents link prediction similarity measures for social networks that exploit the degree distribution of the networks. In the context of link prediction in dense networks, the text proposes similarity measures based on Markov inequality degree thresholding (MIDTs), which only consider nodes whose degree is above a threshold for a possible link. Also presented are similarity measures based on cliques (CNC, AAC, RAC), which assign extra weight between nodes sharing a greater number of cliques. Additionally, a locally adaptive (LA) similarity measure is proposed that assigns different weights to common nodes based on the degree distribution of the local neighborhood and the degree distribution of the network. In the context of link prediction in dense networks, the text introduces a novel two-phase framework that adds edges to the sparse graph to forma boost graph.

• Link Prediction
• Power Law Degree Distribution
• Local Neighborhood
• Recommender Systems
• Graph Mining

• Department of Computer Science, University of Maryland, College Park, USA

  Virinchi Srinivas

• Dept. Computer Sci & Engg,R No: CS310, Indian Institute of Technology Kharagpur, Kharagpur, India

  Pabitra Mitra

Dr. Virinchi Srinivas is a Graduate Research Assistant in the Department of Computer Science at the University of Maryland, College Park, MD, USA.

Dr. Pabitra Mitra is an Associate Professor in the Department of Computer Science and Engineering at the Indian Institute of Technology, Kharagpur, India.

Policies and ethics