CODEWITHSUNDEEP
rsentiment-analysissentimentr
Jasmine B
Jasmine B

Reputation: 53

Can't get update_polarity_table in Sentimentr to update polarity

I'm trying to do a sentiment analysis using hash_sentiment_socal_google in Sentimentr. Looking through the responses, I've noticed that one word responses of "unsure", or "unknown", get an average sentiment score of -.5. And "yes", gets .8. I would like all of them to show up as 0, or neutral.

I don't actually see any of these words in hash_sentiment_socal_google, so I'm not sure why these responses are being assigned sentiment scores. But I just figured I could add to the key with the following code to set to 0:

updated_socal_google <- 
  sentimentr:::update_polarity_table(lexicon::hash_sentiment_socal_google,
  x = data.frame(words = c('yes', 'unsure', 'unknown'),
  polarity = c(0, 0, 0), stringsAsFactors = FALSE))

But after running the code below:

sentiments_new <- sentiment_by(text_sentences, by = NULL, 
                           averaging.function = average_mean,  
                           updated_socal_google, amplifier.weight = .5, 
                           n.before = 10, n.after = 4)

These one word responses are still getting assigned the same average sentiment scores as before, not 0. Can someone explain what I'm doing wrong?

Thank you!

Upvotes: 0

Views: 50

Answers (1)

Jasmine B
Jasmine B

Reputation: 53

Found out the answer, so wanted to update in case anyone else runs into this issue. I needed to specify that polarity_dt = updated_socal_google.

So instead of what I had above:

sentiments_new <- sentiment_by(text_sentences, by = NULL, 
                           averaging.function = average_mean,  
                           polarity_dt = updated_socal_google, amplifier.weight = .5, 
                           n.before = 10, n.after = 4)

Upvotes: 0

Related Questions