Classification of Literary Documents Using the Choquet Integral
Résumé
In literature-oriented Digital Humanities (DH), classifying documents into empirically observed categories can help researchers better understand those concepts they are studying. In machine learning terms, this task corresponds to a classification problem. We address this task with the following characteristics: 1) the considered features are the texts metadata, e.g. author, title and publication date, in addition to the text content; 2) the features are considered both individually and through their coalition, i.e when several features can reinforce each other to influence the prediction. 3) the considered concepts correspond to the author’s gender or to more complex
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| licence |
