ankitasharma-rgb / sentiment-analysis-on-hindi-reviews Goto Github PK
View Code? Open in Web Editor NEWThis project forked from shubham721/sentiment-analysis-on-hindi-reviews
We have used 250 sentences of movie reviews available for research from IIT bombay and also crawled and manually annotated 750 reviews from jagran.com, In total 1000 reviews. After preprocessing the dataset, We generate the featureset as a vector-based approach using Term frequency, tfidf for unigrams and bigrams. Then we used three approaches to predict the sentiment of a review. Approaches used are Resource based, In-language semantic analysis and Machine Translation based semantic analysis.