RankMerging

a supervised learning-to-rank framework to predict links in large social networks

Research output: Contribution to journalArticle

Abstract

Uncovering unknown or missing links in social networks is a difficult task because of their sparsity and because links may represent different types of relationships, characterized by different structural patterns. In this paper, we define a simple yet efficient supervised learning-to-rank framework, called RankMerging, which aims at combining information provided by various unsupervised rankings. We illustrate our method on three different kinds of social networks and show that it substantially improves the performances of unsupervised methods of ranking as well as standard supervised combination strategies. We also describe various properties of RankMerging, such as its computational complexity, its robustness to feature selection and parameter estimation and discuss its area of relevance: the prediction of an adjustable number of links on large networks.

Original languageEnglish
Pages (from-to)1729-1756
Number of pages28
JournalMachine Learning
Volume108
Issue number10
DOIs
Publication statusPublished - 14 Oct 2019

Fingerprint

Supervised learning
Robustness (control systems)
Parameter estimation
Feature extraction
Computational complexity

Keywords

  • Large networks
  • Learning to rank
  • Link prediction
  • Social network analysis
  • Supervised learning

Cite this

@article{77cd5761620145f0b1f2713df42d8b0f,
title = "RankMerging: a supervised learning-to-rank framework to predict links in large social networks",
abstract = "Uncovering unknown or missing links in social networks is a difficult task because of their sparsity and because links may represent different types of relationships, characterized by different structural patterns. In this paper, we define a simple yet efficient supervised learning-to-rank framework, called RankMerging, which aims at combining information provided by various unsupervised rankings. We illustrate our method on three different kinds of social networks and show that it substantially improves the performances of unsupervised methods of ranking as well as standard supervised combination strategies. We also describe various properties of RankMerging, such as its computational complexity, its robustness to feature selection and parameter estimation and discuss its area of relevance: the prediction of an adjustable number of links on large networks.",
keywords = "Large networks, Learning to rank, Link prediction, Social network analysis, Supervised learning",
author = "Lionel Tabourier and Bernardes, {Daniel F.} and Libert, {Anne Sophie} and Renaud Lambiotte",
year = "2019",
month = "10",
day = "14",
doi = "10.1007/s10994-019-05792-4",
language = "English",
volume = "108",
pages = "1729--1756",
journal = "Machine Learning",
issn = "0885-6125",
publisher = "Springer New York",
number = "10",

}

RankMerging : a supervised learning-to-rank framework to predict links in large social networks. / Tabourier, Lionel; Bernardes, Daniel F.; Libert, Anne Sophie; Lambiotte, Renaud.

In: Machine Learning, Vol. 108, No. 10, 14.10.2019, p. 1729-1756.

Research output: Contribution to journalArticle

TY - JOUR

T1 - RankMerging

T2 - a supervised learning-to-rank framework to predict links in large social networks

AU - Tabourier, Lionel

AU - Bernardes, Daniel F.

AU - Libert, Anne Sophie

AU - Lambiotte, Renaud

PY - 2019/10/14

Y1 - 2019/10/14

N2 - Uncovering unknown or missing links in social networks is a difficult task because of their sparsity and because links may represent different types of relationships, characterized by different structural patterns. In this paper, we define a simple yet efficient supervised learning-to-rank framework, called RankMerging, which aims at combining information provided by various unsupervised rankings. We illustrate our method on three different kinds of social networks and show that it substantially improves the performances of unsupervised methods of ranking as well as standard supervised combination strategies. We also describe various properties of RankMerging, such as its computational complexity, its robustness to feature selection and parameter estimation and discuss its area of relevance: the prediction of an adjustable number of links on large networks.

AB - Uncovering unknown or missing links in social networks is a difficult task because of their sparsity and because links may represent different types of relationships, characterized by different structural patterns. In this paper, we define a simple yet efficient supervised learning-to-rank framework, called RankMerging, which aims at combining information provided by various unsupervised rankings. We illustrate our method on three different kinds of social networks and show that it substantially improves the performances of unsupervised methods of ranking as well as standard supervised combination strategies. We also describe various properties of RankMerging, such as its computational complexity, its robustness to feature selection and parameter estimation and discuss its area of relevance: the prediction of an adjustable number of links on large networks.

KW - Large networks

KW - Learning to rank

KW - Link prediction

KW - Social network analysis

KW - Supervised learning

UR - http://www.scopus.com/inward/record.url?scp=85063218991&partnerID=8YFLogxK

U2 - 10.1007/s10994-019-05792-4

DO - 10.1007/s10994-019-05792-4

M3 - Article

VL - 108

SP - 1729

EP - 1756

JO - Machine Learning

JF - Machine Learning

SN - 0885-6125

IS - 10

ER -