Embedding issues

[sunggukcha]

I have a question about the implementation of the experiment. You used predefined word embedding as a classifier, then what did you assigned embedding for background? In other words, there are background class (labeled 0). How did you implement the classifier regarding the classes? Or which word did you use for the class 0? I am comparing your 'fasttext.pkl' file with 'Common Crawl pretrained Fasttext', but I cannot find any matching word.