Opinion Analysis of text on the basis of Three Domain Classification

In today's world, everyone is aware of comment analysis. Its better example is if we want to buy gadgets online or we want to watch a movie, we first see comments or ratings of that product or movie. For this, every website asks every customer to give their feedback which is helpful for product success or movie success. From this feedback or opinion the site owner grabs the mood of the customer. That's why every site owner wants a comment from customer. For searching the sentiment of text which is given by customer there are a number of methods available. In the proposed system the sentiment is finding out in semi-supervised manner. By using this system, polarity or domain classification of text is done. In this system, we combine LDA with data expansion method for data preprocessing which is known as LDA preprocessing and then Naïve Bayes dual training and Domain prediction algorithms are used for sentiment evaluation. From this, we will classify the sentence in three domains i.e. positive, negative and neutral.