Cluster Based Sentiment Analysis in Cross Domain

Sentiment analysis aims to detect attitude and feelings of the opinion holder for the given reviews. Reviews are user generated content about wide variety of commodities available on the social media. It is useful for both the consumer and business people. Sentiment Classifier (SC) might classify reviews as Positive or Negative based on the sentiment expressed in review. SC are domain dependent (same opinion word gives different meaning or contrast polarity in different domain). Cross domain sentiment classification problem were challenges of training a classifier from one or more domain and applying the trained classifier on different domain. The classifier which automatically classifies the reviews as positive sentiments and negative sentiments in different domain is handled. The user reviews from social media is preprocessed for extracting the opinion word from reviews and we classify the word as dependent and independent word using spectral feature alignment by constructing a bipartite graph. After the classified word is check for its polarity and feature based summary.