Sentimental analysis performed on the Amazon book reviews
-
Scrape at least 5000 reviews/blogs from the web using R and then apply the sentiment analysis technique that you have already learnt using a downloaded positive/negative lexicon. Stem both lexicon and reviews using openNLP. There are many R scraping tools such as rvest. Prepare a labeled dataset.
-
Run SVM, DBN, and CNN in R on the labeled review dataset for sentiment classification using deepNet/Darch and MXNet. Your input corresponding to a review is its sparse representation.