I am trying to use a fully custom lexicon (based on frequency table of words for my dataset, with custom values). The format is the same as in the default VADER lexicon. However, I can not get the VADER script to use my custom lexicon.
I have tried changing the filepath in sentiment_analyzer.py, and I have tried replacing the vader_lexicon.txt file in the default path (in appdata/roaming/nltk_data....). Neither option uses my custom lexicon, and instead it still uses the original VADER lexicon.