Sentiment Analysis is one the key Application of Text Analytics. With Vast amount of text data at their dispoal, leading oranigzations wants to utilize the unstructred data to better understand their customer and their requiremnts.
for example Amazon may want to understand customer reviews or Netflix may want to gauge customer interest or may be Political parties want to understand sentiment in their constituencies. There are many more such real life scenario where sentiment Analysis can helps us to better understand our clients.
Though it is still an area of research but below are 3 ways we can implement sentiment analysis-
- Lexicon based dictionary Apporach
- Traditional bag of word model with Tf-idf
- State of art deep learning models (Doc2Vec)
After Implementing all the above techniques on IMDb movie reviews, One can find that Doc2Vec models are really working well for sentmiment Analysis. I've acheived 92% accuracy on test data.