Scalable Privacy-Compliant Virality Prediction on Twitter

Damian Kowalczyk, Jan Larsen

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

177 Downloads (Pure)

Abstract

The digital town hall of Twitter becomes a preferred medium of communication for individuals and organizations across the globe. Some of them reach audiences of millions, while others struggle to get noticed. Given the impact of social media, the question remains more relevant than ever: how to model the dynamics of attention in Twitter. Researchers around the world turn to machine learning to predict the most influential tweets and authors, navigating the volume, velocity, and variety of social big data, with many compromises. In this paper, we revisit content popularity prediction on Twitter. We argue that strict alignment of data acquisition, storage and analysis algorithms is necessary to avoid the common trade-offs between scalability, accuracy and privacy compliance. We propose a new framework for the rapid acquisition of large-scale datasets, high accuracy supervisory signal and multilanguage sentiment prediction while respecting every privacy request applicable. We then apply a novel gradient boosting framework to achieve state-of-the-art results in virality ranking, already before including tweet’s visual or propagation features. Our Gradient Boosted Regression Tree is the first to offer explainable, strong ranking performance on benchmark datasets. Since the analysis focused on features available early, the model is immediately applicable to incoming tweets in 18 languages.
Original languageEnglish
Title of host publicationProceedings of AffCon 2019
PublisherCEUR-WS
Publication date2019
Pages12-27
Publication statusPublished - 2019
EventAAAI-19 WORKSHOP ON AFFECTIVE CONTENT ANALYSIS & CL-AFF HAPPINESS SHARED TASK - Hilton Hawaiian Village, Honolulu, United States
Duration: 27 Jan 201928 Jan 2019

Seminar

SeminarAAAI-19 WORKSHOP ON AFFECTIVE CONTENT ANALYSIS & CL-AFF HAPPINESS SHARED TASK
LocationHilton Hawaiian Village
CountryUnited States
CityHonolulu
Period27/01/201928/01/2019
SeriesCEUR Workshop Proceedings
Volume2328
ISSN1613-0073

Cite this

Kowalczyk, D., & Larsen, J. (2019). Scalable Privacy-Compliant Virality Prediction on Twitter. In Proceedings of AffCon 2019 (pp. 12-27). CEUR-WS. CEUR Workshop Proceedings, Vol.. 2328
Kowalczyk, Damian ; Larsen, Jan. / Scalable Privacy-Compliant Virality Prediction on Twitter. Proceedings of AffCon 2019. CEUR-WS, 2019. pp. 12-27 (CEUR Workshop Proceedings, Vol. 2328).
@inproceedings{b0ba5af712464332aaf6d7eb766c1faa,
title = "Scalable Privacy-Compliant Virality Prediction on Twitter",
abstract = "The digital town hall of Twitter becomes a preferred medium of communication for individuals and organizations across the globe. Some of them reach audiences of millions, while others struggle to get noticed. Given the impact of social media, the question remains more relevant than ever: how to model the dynamics of attention in Twitter. Researchers around the world turn to machine learning to predict the most influential tweets and authors, navigating the volume, velocity, and variety of social big data, with many compromises. In this paper, we revisit content popularity prediction on Twitter. We argue that strict alignment of data acquisition, storage and analysis algorithms is necessary to avoid the common trade-offs between scalability, accuracy and privacy compliance. We propose a new framework for the rapid acquisition of large-scale datasets, high accuracy supervisory signal and multilanguage sentiment prediction while respecting every privacy request applicable. We then apply a novel gradient boosting framework to achieve state-of-the-art results in virality ranking, already before including tweet’s visual or propagation features. Our Gradient Boosted Regression Tree is the first to offer explainable, strong ranking performance on benchmark datasets. Since the analysis focused on features available early, the model is immediately applicable to incoming tweets in 18 languages.",
author = "Damian Kowalczyk and Jan Larsen",
year = "2019",
language = "English",
pages = "12--27",
booktitle = "Proceedings of AffCon 2019",
publisher = "CEUR-WS",

}

Kowalczyk, D & Larsen, J 2019, Scalable Privacy-Compliant Virality Prediction on Twitter. in Proceedings of AffCon 2019. CEUR-WS, CEUR Workshop Proceedings, vol. 2328, pp. 12-27, AAAI-19 WORKSHOP ON AFFECTIVE CONTENT ANALYSIS & CL-AFF HAPPINESS SHARED TASK, Honolulu, United States, 27/01/2019.

Scalable Privacy-Compliant Virality Prediction on Twitter. / Kowalczyk, Damian; Larsen, Jan.

Proceedings of AffCon 2019. CEUR-WS, 2019. p. 12-27 (CEUR Workshop Proceedings, Vol. 2328).

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - Scalable Privacy-Compliant Virality Prediction on Twitter

AU - Kowalczyk, Damian

AU - Larsen, Jan

PY - 2019

Y1 - 2019

N2 - The digital town hall of Twitter becomes a preferred medium of communication for individuals and organizations across the globe. Some of them reach audiences of millions, while others struggle to get noticed. Given the impact of social media, the question remains more relevant than ever: how to model the dynamics of attention in Twitter. Researchers around the world turn to machine learning to predict the most influential tweets and authors, navigating the volume, velocity, and variety of social big data, with many compromises. In this paper, we revisit content popularity prediction on Twitter. We argue that strict alignment of data acquisition, storage and analysis algorithms is necessary to avoid the common trade-offs between scalability, accuracy and privacy compliance. We propose a new framework for the rapid acquisition of large-scale datasets, high accuracy supervisory signal and multilanguage sentiment prediction while respecting every privacy request applicable. We then apply a novel gradient boosting framework to achieve state-of-the-art results in virality ranking, already before including tweet’s visual or propagation features. Our Gradient Boosted Regression Tree is the first to offer explainable, strong ranking performance on benchmark datasets. Since the analysis focused on features available early, the model is immediately applicable to incoming tweets in 18 languages.

AB - The digital town hall of Twitter becomes a preferred medium of communication for individuals and organizations across the globe. Some of them reach audiences of millions, while others struggle to get noticed. Given the impact of social media, the question remains more relevant than ever: how to model the dynamics of attention in Twitter. Researchers around the world turn to machine learning to predict the most influential tweets and authors, navigating the volume, velocity, and variety of social big data, with many compromises. In this paper, we revisit content popularity prediction on Twitter. We argue that strict alignment of data acquisition, storage and analysis algorithms is necessary to avoid the common trade-offs between scalability, accuracy and privacy compliance. We propose a new framework for the rapid acquisition of large-scale datasets, high accuracy supervisory signal and multilanguage sentiment prediction while respecting every privacy request applicable. We then apply a novel gradient boosting framework to achieve state-of-the-art results in virality ranking, already before including tweet’s visual or propagation features. Our Gradient Boosted Regression Tree is the first to offer explainable, strong ranking performance on benchmark datasets. Since the analysis focused on features available early, the model is immediately applicable to incoming tweets in 18 languages.

M3 - Article in proceedings

SP - 12

EP - 27

BT - Proceedings of AffCon 2019

PB - CEUR-WS

ER -

Kowalczyk D, Larsen J. Scalable Privacy-Compliant Virality Prediction on Twitter. In Proceedings of AffCon 2019. CEUR-WS. 2019. p. 12-27. (CEUR Workshop Proceedings, Vol. 2328).