A toolkit for estimating Chinese sentiment scores with multiple measures.
pip install cn-sentiment-measures
- Absolute Proportional Difference - APD. Bounds: [-1, 1]
- Relative Proportional Difference - RPD. Bounds: [-1, 1]
- Logit scale - LC. Bounds: [-infinity, +infinity]
Improved versions of the above measures by integrating degree words are marked with _with_degree
in the toolkit.
from cn_sentiment_measures.sentiment_measures import SentimentMeasures
sm = SentimentMeasures()
word_list=["我","今天","很","高兴"]
print("Here are measure values with normal measures!")
print('APD = ',sm.APD(word_list))
print('RPD = ',sm.RPD(word_list))
print('LS = ',sm.LS(word_list))
print()
print("Here are measure values optimized by using adverb words!")
print('APD* = ',sm.APD_with_degree(word_list))
print('RPD* = ',sm.RPD_with_degree(word_list))
print('LS* = ',sm.LS_with_degree(word_list))
The project is provided by Donghua Chen.