Analyzing Sentiment of Movie Review Data using Naive Bayes Neural Classifier
Lina L. Dhande, Girish Kumar Patnaik · 2014
Now a days sentiment analysis is active field of research, to extract people's opinion about particular product or service. The most useful application of sentiment analysis is the sentiment classification of product reviews. The task of sentiment classification is to classify reviews of user as positive or negative from textual information alone. For that purpose many researchers used data mining classification techniques such as Naive Bayes classifier with strong independence assumption. But Naive Bayes classifier lack in accuracy for many complex real-world situations where there exists dependency among features. Further, the Neural Network with appropriate network structure is good enough to handle the correlation or dependence between input variables. In proposed system, the Naive Bayes and Neural Network classifier are combined for sentiment classification. In Experimental results, the movie review is classified into positive or negative polarities of sentiment using classifiers. The accuracy of sentiment analysis is increased upto 80.65% by combining Naive Bayes classifier with Neural Network for unigram feature on movie review dataset.