RT Journal Article
T1 Interpretable classification of Wiki-review streams
A1 García Méndez, Silvia
A1 Leal, Fátima
A1 Malheiro, Benedita
A1 Burguillo Rial, Juan Carlos
K1 3325 Tecnología de las Telecomunicaciones
K1 6308 Comunicaciones Sociales
K1 1203.17 Informática
AB Wiki articles are created and maintained by a crowd of editors, producing a continuous streamof reviews. Reviews can take the form of additions, reverts, or both. This crowdsourcing model is exposedto manipulation since neither reviews nor editors are automatically screened and purged. To protect articlesagainst vandalism or damage, the stream of reviews can be mined to classify reviews and profle editors inreal-time. The goal of this work is to anticipate and explain which reviews to revert. This way, editors areinformed why their edits will be reverted. The proposed method employs stream-based processing, updatingthe profling and classifcation models on each incoming event. The profling uses side and content-basedfeatures employing Natural Language Processing, and editor profles are incrementally updated based ontheir reviews. Since the proposed method relies on self-explainable classifcation algorithms, it is possibleto understand why a review has been classifed as a revert or a non-revert. In addition, this work contributesan algorithm for generating synthetic data for class balancing, making the fnal classifcation fairer. Theproposed online method was tested with a real data set from Wikivoyage, which was balanced through theaforementioned synthetic data generation. The results attained near-90 % values for all evaluation metrics(accuracy, precision, recall, and F-measure)
PB IEEE Access
SN 21693536
YR 2023
FD 2023-12-13
LK http://hdl.handle.net/11093/6643
UL http://hdl.handle.net/11093/6643
LA eng
NO IEEE Access, 11, 141137-141151 (2023)
NO Fundação para a Ciência e a Tecnologia | Ref. UIDB/50014/2020
DS Investigo
RD 13-sep-2024