Next Article in Journal
Accuracy and Reliability of Local Positioning Systems for Measuring Sport Movement Patterns in Stadium-Scale: A Systematic Review
Next Article in Special Issue
Incorporating Synonym for Lexical Sememe Prediction: An Attention-Based Model
Previous Article in Journal
A Study on Tensile Strain Distribution and Fracture Coordinate of Nanofiber Mat by Digital Image Correlation System
Previous Article in Special Issue
DAWE: A Double Attention-Based Word Embedding Model with Sememe Structure Information
Article

Zero-Shot Learning for Cross-Lingual News Sentiment Classification

1
Jožef Stefan Institute, 1000 Ljubljana, Slovenia
2
Jožef Stefan International Postgraduate School, 1000 Ljubljana, Slovenia
3
Trikoder d.o.o., 10010 Zagreb, Croatia
*
Authors to whom correspondence should be addressed.
Appl. Sci. 2020, 10(17), 5993; https://doi.org/10.3390/app10175993
Received: 31 July 2020 / Revised: 21 August 2020 / Accepted: 25 August 2020 / Published: 29 August 2020
In this paper, we address the task of zero-shot cross-lingual news sentiment classification. Given the annotated dataset of positive, neutral, and negative news in Slovene, the aim is to develop a news classification system that assigns the sentiment category not only to Slovene news, but to news in another language without any training data required. Our system is based on the multilingual BERTmodel, while we test different approaches for handling long documents and propose a novel technique for sentiment enrichment of the BERT model as an intermediate training step. With the proposed approach, we achieve state-of-the-art performance on the sentiment analysis task on Slovenian news. We evaluate the zero-shot cross-lingual capabilities of our system on a novel news sentiment test set in Croatian. The results show that the cross-lingual approach also largely outperforms the majority classifier, as well as all settings without sentiment enrichment in pre-training. View Full-Text
Keywords: sentiment analysis; zero-shot learning; news analysis; cross-lingual classification; multilingual transformers sentiment analysis; zero-shot learning; news analysis; cross-lingual classification; multilingual transformers
Show Figures

Figure 1

MDPI and ACS Style

Pelicon, A.; Pranjić, M.; Miljković, D.; Škrlj, B.; Pollak, S. Zero-Shot Learning for Cross-Lingual News Sentiment Classification. Appl. Sci. 2020, 10, 5993. https://doi.org/10.3390/app10175993

AMA Style

Pelicon A, Pranjić M, Miljković D, Škrlj B, Pollak S. Zero-Shot Learning for Cross-Lingual News Sentiment Classification. Applied Sciences. 2020; 10(17):5993. https://doi.org/10.3390/app10175993

Chicago/Turabian Style

Pelicon, Andraž; Pranjić, Marko; Miljković, Dragana; Škrlj, Blaž; Pollak, Senja. 2020. "Zero-Shot Learning for Cross-Lingual News Sentiment Classification" Appl. Sci. 10, no. 17: 5993. https://doi.org/10.3390/app10175993

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Search more from Scilit
 
Search
Back to TopTop