Is there a sentiment words list? It will be more useful if the words are classified into the sub sentiment values (-ve, +ve, very +ve, very -ve, etc)?
Hi @ahmed_gomaa -
In some of the example workflows (like this one, which uses a lexicon based approach), we use positive and negative sentiment words from the MPQA Corpus (https://mpqa.cs.pitt.edu/corpora/mpqa_corpus/). But this is not built into a node by default.
2 Likes
Yes, I note that from the labs. So, I should build my own list for each one of them, shouldn’t I?
Why isn’t there a built-in list for that type of tagging like as the POS tagging, NE tagging, or stop word filter?
As for why, I’m not exactly sure. If that would be useful for you, we could consider implementing something like a Sentiment Assignment node.
(Note that we recently released the Amazon Comprehend node, which uses an external Amazon service to classify text according to sentiment.)
1 Like
Thanks. I will try it
How to build file readers based on the mpqa input of positive and and bad words
Take a look at the workflow below, which shows how to read the MPQA lists using a couple of File Reader nodes.
2 Likes
I don’t know if this is what u are looking for but I’ll attach down below two CSVs, one with a positive list of words and another with the negative ones:
https://drive.google.com/drive/folders/1d1C7ITDXuMSYkZAINuiZnq9Z8Iv-OpCb?usp=sharing
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.