cjhutto / vaderSentiment

VADER Sentiment Analysis. VADER (Valence Aware Dictionary and sEntiment Reasoner) is a lexicon and rule-based sentiment analysis tool that is specifically attuned to sentiments expressed in social media, and works well on texts from other domains.
MIT License
4.43k stars 1k forks source link

Is it better to use pos, neu and neg to determine the sentiment? #92

Closed bright1993ff66 closed 4 years ago

bright1993ff66 commented 4 years ago

Hi, I am working on an twitter sentiment classification task. I think the compound method does not work well in cases like the following:

text: "I am at ping chi street bus stop."

The result is like the following:

{'pos': 0.0, 'neg': 0.268, 'compound': -0.296, 'neu': 0.732}

The stop here just means the bus stop. But the sentiment analyzer regards this as a negative notation...

cjhutto commented 4 years ago

Thanks for identifying this.
Made a change that accounts for "bus stop", and is extensible as new cases are identified.