Influential tweeters in relation to highly cited articles in altmetric big data

Saeed-Ul Hassan (), Timothy D. Bowman (), Mudassir Shabbir, Aqsa Akhtar, Mubashir Imran and Naif Radi Aljohani ()
Additional contact information
Saeed-Ul Hassan: Information Technology University
Timothy D. Bowman: Wayne State University
Mudassir Shabbir: Information Technology University
Aqsa Akhtar: Information Technology University
Mubashir Imran: Information Technology University
Naif Radi Aljohani: King Abdulaziz University

Scientometrics, 2019, vol. 119, issue 1, No 22, 493 pages

Abstract: Abstract The relationship between influential tweeters and highly cited articles in the field of information sciences was analysed using Twitter data gathered by Altmetric.com from July 2011 through February 2017. The dataset consists of more than 10,000 tweets, and these mentions, retweets and followers were used to generate a connected, undirected graph. This graph reveals the most influential tweeters by identifying the largest drop in the eigenvalue of adjacency or affinity matrix of a graph when certain nodes are removed; those which, when deleted, cause the greatest drop in the eigenvalue of the graph are considered to be the most influential. The machine-learning model applied in this work utilizes a feature vector containing the accumulated sum of the rank scores of those influential users who tweet a given article, along with known altmetric features such as the user type and post counts for various social media. Finally, the supervised-learning model was trained using Random Forest and Support Vector Machine classifiers with 11 features, including the sum of the ranks of influential users who tweet a given article in our dataset. The results were analysed using Receiver Operating Characteristic (ROC) curves and Precision Recall (PR) curves, which give the commendable outcomes compared to the baseline model. We found that, for the classification of highly cited articles, Twitter users’ score for influence is the most important feature. Finally, we show that our model—which was trained by taking the score for influence into consideration—outperforms the baseline, at 79% for ROC and 90% for PR with the Random Forest Model, effectively identifying the highly cited articles.

Keywords: Altmetrics; Influential users; Twitter; Highly cited articles (search for similar items in EconPapers)
Date: 2019
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (5)

Downloads: (external link)
http://link.springer.com/10.1007/s11192-019-03044-9 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:scient:v:119:y:2019:i:1:d:10.1007_s11192-019-03044-9

Ordering information: This journal article can be ordered from
http://www.springer.com/economics/journal/11192

DOI: 10.1007/s11192-019-03044-9

Access Statistics for this article

Scientometrics is currently edited by Wolfgang Glänzel

More articles in Scientometrics from Springer, Akadémiai Kiadó
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().