Sentiment analysis

sentiment_analysis

#1

H all,

Generally for a sentiment analysis we count the number of +ve or -ve words in a review and which ever is higher we give that sentiment to that review. Is there any other way which can actually detect sarcasm or fallow the semantic meaning to find sentiment???


#2

@raviteja1993 there is a complete hands on tutorial on sentiment analysis. Do check it out.


#3

thank you


#4

@pjoshi15 In the above mentioned article you have removed labels from the train ,combined train and test sets, then build a DTM , then again divided the dataset proportionally as before into train and test. Now you again added label to this new DTM train and build the model and predicted the test.

Now If I give you new sentence with words that are present in train or test then how this model will predict its sentiment.