Ταξινόμηση κειμένου πολιτικού λόγου με τεχνολογίες BERT
Text classification of political speech with BERT
View/
Keywords
Επεξεργασία φυσικής γλώσσας ; Διαδική ταξινόμηση ; Βαθιά μάθηση ; BERT ; Κλάσεις δημοκρατικού - ρεπουμπλικανικού λόγουAbstract
Text classification, as a natural language processing task, has been a popular and interesting field
of Machine Learning. Over the years, there has been a significant advancement in how the
analysis of complex texts can be conducted with the utilization of machine and deep learning
networks.
This thesis aims to develop a classifier for texts, coming from American political speech, capable
of performing binary classification to the classes of democratic or republican speech.
BERT deep learning algorithms are ideal models that incorporate linguistic capabilities. With their
architecture based on transformers networks, they have effectiveness in processing natural
language. To achieve our goal, we adapt a model from this family to the text data we have
collected. By conducting two independent experiments with different parameters, the success of
this process is tested.
Department
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
55Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού - Παρόμοια Διανομή 3.0 Ελλάδα
Αναφορά Δημιουργού - Παρόμοια Διανομή 3.0 Ελλάδα