Sentiment Analysis - Is there a dictionary of “hate” words and similarly for “compassionate/neutral” words?

nlp
sentiment_analysis

#1

Hi All,
This question is about how we should go about building a lexicon of words which can be used for scoring the documents in a sentiment analysis problem. The scoring could be required to categorize a sentence/document as “Positive”, “Negative”, “Hate”, “Compassionate” etc based on the problem at hand.
The Jeffrey-Been algorithm helps in scoring sentences as having positive and/or negative sentiment. The dictionaries are publicly available to use for similar problems.
However if we have a different problem statement we need to have a different set of lexicons. How do we create them or do we have any such lexicons available in any site.
I was working on sentiment analysis problem in AV. Wondering if you could help me…
We are required to identify the tweets which are hate tweets and which are not.
I am planning to use lexicon based approach to make the algorithm learn the sentiment type and then use it to predict.
I am needing some assistance in being able to create a dictionary of “hate” words and similarly for “compassionate/neutral” words.

Regards,
Pranov


#2

You could try using Stanford CoreNLP package. It has several tools to extract sentiment as a value which is negative for hate words. If you do have the complete list of words, you could just use the annotate function to extract ‘negativity score’ for each word.


#3

Thanks for the suggestion. I will look for the package and try to use it.
could you share a link which will help to create the bag of hate words…


#4

Hi Pranov,

I recommend to use vader python package which gives sentiment score of each word using their pre existing dictionary. You can look into that and it also performs some rule based analysis to score a sentence.