Word Embedding für Produktreviews

Marc Schulder, Michael Wiegand
27. November 2017
Datensatz Projekt DOI Publikation
Diese Resource ist nur auf Englisch verfügbar.

DOI licence: CC BY 4.0

A word embedding of the Amazon Product Review Corpus (Jindal and Liu, 2008).

Created using Word2Vec in CBOW mode, 500 dimensions and window size 5.

Words have been lemmatised and particle verbs have been merged into a single token (e.g. calm_down).

Attribution

This dataset was created as part of our IJCNLP 2017 paper “Towards Bootstrapping a Polarity Shifter Lexicon using Linguistic Features”. If you use the dataset in your research or work, please cite the publication.