Trend reservoir detection: Minimal persistence and resonant behavior of trends in social media

TY - GEN

T1 - Trend reservoir detection

T2 - 1st Workshop on Computational Humanities Research, CHR 2020

AU - Nielbo, Kristoffer L.

AU - Vahlstrup, Peter B.

AU - Bechmann, Anja

AU - Gao, Jianbo

N1 - Publisher Copyright: © 2020 Copyright for this paper by its authors. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.

PY - 2020

Y1 - 2020

N2 - Sociocultural trends from social media platforms such as Twitter or Instagram have become an important part of knowledge discovery. The 'trend' construct is however ambiguous and its estimation from unstructured sociocultural data complicated by several methodological issues. This paper presents an approach to trend estimation that combines domain knowledge of social media with advances in information theory and dynamical systems. In particular, we show how trend reservoirs (i.e., signals that display trend potential) can be identified by their relationship between novel and resonant behavior, and their minimal persistence.This approach contrasts with trend estimation that relies on linear or polynomial techniques to study point-like novelty behavior in social media, and it completes approaches that rely on smooth functions of time.

AB - Sociocultural trends from social media platforms such as Twitter or Instagram have become an important part of knowledge discovery. The 'trend' construct is however ambiguous and its estimation from unstructured sociocultural data complicated by several methodological issues. This paper presents an approach to trend estimation that combines domain knowledge of social media with advances in information theory and dynamical systems. In particular, we show how trend reservoirs (i.e., signals that display trend potential) can be identified by their relationship between novel and resonant behavior, and their minimal persistence.This approach contrasts with trend estimation that relies on linear or polynomial techniques to study point-like novelty behavior in social media, and it completes approaches that rely on smooth functions of time.

KW - Fractal analysis

KW - Information dynamics

KW - Social media

KW - Trend detection

UR - http://www.scopus.com/inward/record.url?scp=85095965082&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85095965082

SN - 1613-0073

VL - 2723

SP - 290

EP - 297

JO - CEUR Workshop Proceedings

JF - CEUR Workshop Proceedings

Y2 - 18 November 2020 through 20 November 2020

ER -