Pronounced "sen-T-M".
senTM is a lexical approach to sentiment analysis in Danish. The approach is inspired by the afinn package, but using the Danish Sentiment Lexicon version 0.2 (2022-12-20).
The approach use the part-of-speech (POS) model from Spacy and then matches any tokens from the Danish Sentiment Lexicon.
License: Same as Danish Sentiment Lexicon version 0.2 (2022-12-20): CC-BY-SA 4.0 International https://creativecommons.org/licenses/by-sa/4.0/
Install from pip:
pip install sentm
First initialize model.
from sentm.sentm import senTM
sentm_model = senTM()
You can both get sentiment score:
sentm_model.score('Du er en kæmpe idiot!')
You can also use it as a classifier. Here, the labels are determined by:
- Score larger than 1: "positiv"
- Score between -1 and 1: "neutral"
- Score lower than -1: "negativ"
sentm_model.classify('Du er en kæmpe idiot!')