Analysis and Prediction of Dyads in Twitter

Isa Inuwa-Dutse, Mark Liptrott, Yannis Korkontzelos

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

Social networks are useful for linking micro and macro levels of sociological theory by enabling the analysis of various forms of relationships. In social science, a taxonomy of social relationships is described as a function of closeness among users. The closer the users are, the more cohesive and trustworthy. Identifying dyadic ties, pairs of fully connected users, on Twitter is challenging due to the flexible and eccentric underlying connection patterns. The ability to follow anyone results in many unidirectional connections between socially disconnected users and ultimately affects clustering users and, in turn, the veracity of online content. Major challenges towards effective user clustering are the low number of dyads and efficient methods to identify more. In this study, we query over 17M verified and unverified Twitter user accounts and retrieve dyadic ties. In the collected data, 55 % and 21 % of unverified and verified profiles, respectively, participate in dyadic ties. We describe the importance of dyads in the detection of cohesive user groups and how they may be used to validate trustworthiness. We demonstrate how identifying and using dyadic ties will improve Twitter analysis, in the future. Finally, we develop a deep learning model for dyad prediction.

Original languageEnglish
Title of host publicationNatural Language Processing and Information Systems
Subtitle of host publication24th International Conference on Applications of Natural Language to Information Systems, NLDB 2019, Salford, UK, June 26–28, 2019, Proceedings
EditorsElisabeth Métais, Farid Meziane, Sunil Vadera, Vijayan Sugumaran, Mohamad Saraee
Place of PublicationCham
PublisherSpringer Nature Switzerland AG
Pages303-311
Number of pages9
Volume11608 LNCS
Edition1st
ISBN (Electronic)9783030232818
ISBN (Print)9783030232801
DOIs
Publication statusPublished - 21 Jun 2019
Externally publishedYes
Event24th International Conference on Application of Natural Language to Information Systems - Salford, United Kingdom
Duration: 26 Jun 201928 Jun 2019
Conference number: 24

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11608 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference24th International Conference on Application of Natural Language to Information Systems
Abbreviated titleNLDB 2019
CountryUnited Kingdom
CitySalford
Period26/06/1928/06/19

Fingerprint

Dive into the research topics of 'Analysis and Prediction of Dyads in Twitter'. Together they form a unique fingerprint.

Cite this