Multi-label sentiment analysis on 100 languages with dynamic weighting for label imbalance
buir.contributor.author | Yılmaz, Selim Fırat | |
buir.contributor.author | Kaynak, Ergün Batuhan | |
buir.contributor.author | Koç, Aykut | |
buir.contributor.author | Dibeklioğlu, Hamdi | |
buir.contributor.author | Kozat, Süleyman Serdar | |
buir.contributor.orcid | Yılmaz, Selim Fırat|0000-0002-0486-7731 | |
buir.contributor.orcid | Kaynak, Ergün Batuhan|0000-0002-3249-3343 | |
buir.contributor.orcid | Koç, Aykut|0000-0002-6348-2663 | |
buir.contributor.orcid | Dibeklioğlu, Hamdi|0000-0003-0851-7808 | |
buir.contributor.orcid | Kozat, Süleyman Serdar|0000-0002-6488-3848 | |
dc.citation.epage | 343 | en_US |
dc.citation.issueNumber | 1 | |
dc.citation.spage | 331 | en_US |
dc.citation.volumeNumber | 34 | |
dc.contributor.author | Yılmaz, Selim Fırat | |
dc.contributor.author | Kaynak, Ergün Batuhan | |
dc.contributor.author | Koç, Aykut | |
dc.contributor.author | Dibeklioğlu, Hamdi | |
dc.contributor.author | Kozat, Süleyman Serdar | |
dc.date.accessioned | 2022-03-04T08:53:25Z | |
dc.date.available | 2022-03-04T08:53:25Z | |
dc.date.issued | 2021-07-19 | |
dc.department | Department of Computer Engineering | en_US |
dc.department | Department of Electrical and Electronics Engineering | en_US |
dc.department | National Magnetic Resonance Research Center (UMRAM) | en_US |
dc.description.abstract | We investigate cross-lingual sentiment analysis, which has attracted significant attention due to its applications in various areas including market research, politics, and social sciences. In particular, we introduce a sentiment analysis framework in multi-label setting as it obeys Plutchik's wheel of emotions. We introduce a novel dynamic weighting method that balances the contribution from each class during training, unlike previous static weighting methods that assign non-changing weights based on their class frequency. Moreover, we adapt the focal loss that favors harder instances from single-label object recognition literature to our multi-label setting. Furthermore, we derive a method to choose optimal class-specific thresholds that maximize the macro-f1 score in linear time complexity. Through an extensive set of experiments, we show that our method obtains the state-of-the-art performance in seven of nine metrics in three different languages using a single model compared with the common baselines and the best performing methods in the SemEval competition. We publicly share our code for our model, which can perform sentiment analysis in 100 languages, to facilitate further research. | en_US |
dc.description.provenance | Submitted by Dilan Ayverdi (dilan.ayverdi@bilkent.edu.tr) on 2022-03-04T08:53:24Z No. of bitstreams: 1 Multi-label_sentiment_analysis_on_100_languages_with_dynamic_weighting_for_label_imbalance.pdf: 1714127 bytes, checksum: a19fa2dd30febece0e044a2d58228885 (MD5) | en |
dc.description.provenance | Made available in DSpace on 2022-03-04T08:53:25Z (GMT). No. of bitstreams: 1 Multi-label_sentiment_analysis_on_100_languages_with_dynamic_weighting_for_label_imbalance.pdf: 1714127 bytes, checksum: a19fa2dd30febece0e044a2d58228885 (MD5) Previous issue date: 2021-07-19 | en |
dc.identifier.doi | 10.1109/TNNLS.2021.3094304 | en_US |
dc.identifier.eissn | 2162-2388 | en_US |
dc.identifier.issn | 2162-237X | en_US |
dc.identifier.uri | http://hdl.handle.net/11693/77683 | en_US |
dc.language.iso | English | en_US |
dc.publisher | Institute of Electrical and Electronics Engineers | en_US |
dc.relation.isversionof | https://doi.org/10.1109/TNNLS.2021.3094304 | en_US |
dc.source.title | IEEE Transactions on Neural Networks and Learning Systems | en_US |
dc.subject | Cross-lingual | en_US |
dc.subject | Label imbalance | en_US |
dc.subject | Macro-f1 maximization | en_US |
dc.subject | Multi-label | en_US |
dc.subject | Natural language processing (NLP) | en_US |
dc.subject | Sentiment analysis | en_US |
dc.subject | Social media | en_US |
dc.title | Multi-label sentiment analysis on 100 languages with dynamic weighting for label imbalance | en_US |
dc.type | Article | en_US |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- Multi-label_sentiment_analysis_on_100_languages_with_dynamic_weighting_for_label_imbalance.pdf
- Size:
- 1.44 MB
- Format:
- Adobe Portable Document Format