Authorship Attribution on Short Texts in the Slovenian Language
The study investigates the task of authorship attribution on short texts in Slovenian using the BERT language model. Authorship attribution is the task of attributing a written text to its author, frequently using stylometry or computational techniques. We create five custom datasets for different n...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/19/10965 |
| _version_ | 1797576184138039296 |
|---|---|
| author | Gregor Gabrovšek Peter Peer Žiga Emeršič Borut Batagelj |
| author_facet | Gregor Gabrovšek Peter Peer Žiga Emeršič Borut Batagelj |
| author_sort | Gregor Gabrovšek |
| collection | DOAJ |
| description | The study investigates the task of authorship attribution on short texts in Slovenian using the BERT language model. Authorship attribution is the task of attributing a written text to its author, frequently using stylometry or computational techniques. We create five custom datasets for different numbers of included text authors and fine-tune two BERT models, SloBERTa and BERT Multilingual (mBERT), to evaluate their performance in closed-class and open-class problems with varying numbers of authors. Our models achieved an F1 score of approximately <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.95</mn></mrow></semantics></math></inline-formula> when using the dataset with the comments of the top five users by the number of written comments. Training on datasets that include comments written by an increasing number of people results in models with a gradually decreasing F1 score. Including out-of-class comments in the evaluation decreases the F1 score by approximately <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.05</mn></mrow></semantics></math></inline-formula>. The study demonstrates the feasibility of using BERT models for authorship attribution in short texts in the Slovenian language. |
| first_indexed | 2024-03-10T21:48:39Z |
| format | Article |
| id | doaj.art-7608638ee2e14dcf949e6581356c8ee0 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T21:48:39Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-7608638ee2e14dcf949e6581356c8ee02023-11-19T14:06:42ZengMDPI AGApplied Sciences2076-34172023-10-0113191096510.3390/app131910965Authorship Attribution on Short Texts in the Slovenian LanguageGregor Gabrovšek0Peter Peer1Žiga Emeršič2Borut Batagelj3Faculty of Computer and Information Science, University of Ljubljana, SI-1000 Ljubljana, SloveniaFaculty of Computer and Information Science, University of Ljubljana, SI-1000 Ljubljana, SloveniaFaculty of Computer and Information Science, University of Ljubljana, SI-1000 Ljubljana, SloveniaFaculty of Computer and Information Science, University of Ljubljana, SI-1000 Ljubljana, SloveniaThe study investigates the task of authorship attribution on short texts in Slovenian using the BERT language model. Authorship attribution is the task of attributing a written text to its author, frequently using stylometry or computational techniques. We create five custom datasets for different numbers of included text authors and fine-tune two BERT models, SloBERTa and BERT Multilingual (mBERT), to evaluate their performance in closed-class and open-class problems with varying numbers of authors. Our models achieved an F1 score of approximately <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.95</mn></mrow></semantics></math></inline-formula> when using the dataset with the comments of the top five users by the number of written comments. Training on datasets that include comments written by an increasing number of people results in models with a gradually decreasing F1 score. Including out-of-class comments in the evaluation decreases the F1 score by approximately <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.05</mn></mrow></semantics></math></inline-formula>. The study demonstrates the feasibility of using BERT models for authorship attribution in short texts in the Slovenian language.https://www.mdpi.com/2076-3417/13/19/10965authorship attributionBERT model fine-tuningdataset construction |
| spellingShingle | Gregor Gabrovšek Peter Peer Žiga Emeršič Borut Batagelj Authorship Attribution on Short Texts in the Slovenian Language Applied Sciences authorship attribution BERT model fine-tuning dataset construction |
| title | Authorship Attribution on Short Texts in the Slovenian Language |
| title_full | Authorship Attribution on Short Texts in the Slovenian Language |
| title_fullStr | Authorship Attribution on Short Texts in the Slovenian Language |
| title_full_unstemmed | Authorship Attribution on Short Texts in the Slovenian Language |
| title_short | Authorship Attribution on Short Texts in the Slovenian Language |
| title_sort | authorship attribution on short texts in the slovenian language |
| topic | authorship attribution BERT model fine-tuning dataset construction |
| url | https://www.mdpi.com/2076-3417/13/19/10965 |
| work_keys_str_mv | AT gregorgabrovsek authorshipattributiononshorttextsintheslovenianlanguage AT peterpeer authorshipattributiononshorttextsintheslovenianlanguage AT zigaemersic authorshipattributiononshorttextsintheslovenianlanguage AT borutbatagelj authorshipattributiononshorttextsintheslovenianlanguage |