VADER( Valence Aware Dictionary for sEntiment Reasoning ) is model used for text sentiment analysis that is sensitive to positive and negative polarity and intensity of emotion.
VADER is available with NLTK package and can be applied directly to unlabeled text data. VADER sentiment analysis relies on dictionary which maps lexical features to emotions intensities called sentiment scores. The sentiment score of text can be obtained by summing up the intensity of each word in text.
For example , words like ‘love’ , ‘like’ , ‘happy’ all convey positive sentiment. but ‘hate’ , ‘ignore’, ‘worst’ convey negative sentiments. VADER is intelligent enough to understand basic context of these words such as ‘did not love’ as negative sentiments.
Let’s explore VADER Sentiment Analysis with NLTK and python
First of all we will import nltk library and download vader_lexicon data set and create object for SentimentIntensityAnalyzer
After that we have loaded review from csv file for amazon reviews and used VADER analysis to get positive or negative analysis. We compared manually given labels with VADER labels and found accuracy score around 71%.
CSV file can be download here : amazonreviews.tsv