[Corpora-List] Influence of word embeddings on classification results

Alexander Osherenko osherenko at gmx.de
Sun Aug 9 14:33:55 CEST 2020


Hello,

I am wondering if there is evidence that word embeddings using fasttext or word2vec significantly improve classification results compared with simple BOW. I am aware that word embeddings consider the context but maybe I can simply extract BOW and use SVM or NB and classification results are acceptable or better.

Best, Alexander

-- Alexander Osherenko, Dr. rer. nat. Ibarus <http://ibarus.org> R&D

Socioware Development <http://www.socioware.de/osherenko_page.html> Humboldt-Universitšt zu Berlin <https://wirsindhumboldt.de/de/VKkZNyFaeu> Senior HCI architect Founder and R&D

Profile: ResearchGate <https://www.researchgate.net/profile/Alexander_Osherenko> Simulating An Intercultural Deontic Scenario Of Urban Life In An Agent-Based Robotic System <https://www.researchgate.net/publication/343223883_Simulating_An_Intercultural_Deontic_Scenario_Of_Urban_Life_In_An_Agent-Based_Robotic_System> -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 2283 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20200809/e6d1b6cd/attachment.txt>



More information about the Corpora mailing list