I am trying to do sentiment analysis with python.I have gone through various tutorials and have used libraries like nltk, textblob etc for it.
But what I want is bit different and I am not able figure out any material for that
Suppose I have a statement like
apples are tasty but they are very expensive
The above statement can be classified in to two classes/labels like taste and money
My aim is to get sentiment of the statement with respect to these two labels
My expected result would be positive sentiment for taste but negative sentiment for money
How this can be achieved
With textblob
def calculate_sentiment_textblob(current_comment):
current_comment = str(current_comment)
comment_sentiment_calculation = TextBlob(current_comment)
comment_sentiment = ""
if comment_sentiment_calculation.sentiment.polarity < 0:
comment_sentiment = "Negative"
elif comment_sentiment_calculation.sentiment.polarity > 0:
comment_sentiment = "Positive"
else:
comment_sentiment = "Neutral"
print(current_comment)
print(comment_sentiment)
sentiment_list.append(current_comment +" "+comment_sentiment)
comments_scraped.loc[comments_scraped.reviews== current_comment,'sentiment_textblob'] = comment_sentiment
With vader
def calculate_sentiment_vader(current_comment):
current_comment = str(current_comment)
comment_sentiment_calculation = sid.polarity_scores(current_comment)
comment_sentiment = ""
if comment_sentiment_calculation['compound'] < 0:
comment_sentiment = "Negative"
elif comment_sentiment_calculation['compound'] > 0:
comment_sentiment = "Positive"
else:
comment_sentiment = "Neutral"
comments_scraped.loc[comments_scraped.reviews== current_comment,'sentiment_vader'] = comment_sentiment
I kindly suggest you investigate on aspect-based sentiment analysis. It does not focus sentiment on only an entity but an entity's attributes. There have been SemEval challenges to investigate this problem on attributes of entities, for example laptops and restaurants.
There were many participants, their papers are published and the organizers published explanatory papers as well.
You can reach them here:
Hope these help, cheers.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.