Multidimensional Author Profling for Social Business Intelligence
Impacte
Scholar |
Altres documents de l'autoria: Lanza Cruz, Indira Lázara; Berlanga Llavori, Rafael; Aramburu Cabo, María José
Metadades
Mostra el registre complet de l'elementcomunitat-uji-handle:10234/9
comunitat-uji-handle2:10234/7038
comunitat-uji-handle3:10234/8634
comunitat-uji-handle4:
INVESTIGACIONMetadades
Títol
Multidimensional Author Profling for Social Business IntelligenceData de publicació
2023Editor
SpringerCita bibliogràfica
Lanza-Cruz, I., Berlanga, R. & Aramburu, M.J. Multidimensional Author Profiling for Social Business Intelligence. Inf Syst Front 26, 195–215 (2024). https://doi.org/10.1007/s10796-023-10370-0Tipus de document
info:eu-repo/semantics/articleVersió de l'editorial
https://link.springer.com/article/10.1007/s10796-023-10370-0Versió
info:eu-repo/semantics/publishedVersionParaules clau / Matèries
Resum
This paper presents a novel author profling method specially aimed at classifying social network users into the multidimensional perspectives for social business intelligence (SBI) applications. In this scenario, being ... [+]
This paper presents a novel author profling method specially aimed at classifying social network users into the multidimensional perspectives for social business intelligence (SBI) applications. In this scenario, being the user profles defned on
demand for each particular SBI application, we cannot assume the existence of labelled datasets for training purposes. Thus,
we propose an unsupervised method to obtain the required labelled datasets for training the profle classifers. Contrary to
other author profling approaches in the literature, we only make use of the users’ descriptions, which are usually part of
the metadata posts. We exhaustively evaluated the proposed method under four diferent tasks for multidimensional author
profling along with state-of-the-art text classifers. We achieved performances around 88% and 98% of F1 score for a gold
standard and a silver standard datasets respectively. Additionally, we compare our results to other supervised approaches
previously proposed for two of our tasks, getting very close performances despite using an unsupervised method. To the best
of our knowledge, this is the frst method designed to label user profles in an unsupervised way for training profle classifers
with a similar performance to fully supervised ones. [-]
Publicat a
Information Systems Frontiers, 2024, 26.Entitat finançadora
Ministerio de Industria y Comercio | Universitat Jaume I
Codi del projecte o subvenció
PDC2021- 121097-I00 | PREDOC/2017/28
Drets d'accés
© The Author(s) 2023
info:eu-repo/semantics/openAccess
info:eu-repo/semantics/openAccess
Apareix a les col.leccions
- ICC_Articles [423]
- LSI_Articles [362]