An Approach to Perform Sentiment Analysis using Data Mining Algorithms

Abstract

To perform the sentiment analysis as a basis for defining and extracting subjective information from sources or easily relating to the identification phase of the polarity of the text, the concept of Natural Processing is used. Participatory approach is required to perform this analysis. It was also called opinion mining as it extracts a user's view or perspective. There are many attributes which pose a problem with knowledge. It is an arbitrary for choosing assets giving a wider range of values. In the current paper, various algorithms of classification are used and it is concluded that the best algorithm is random forest. The issue is that decision trees, especially if the tree is particularly deep, are vulnerable to being over fit. To minimize the bias and error of variance, classification along with random forest classification is used. Through practicing on different data sets, random forests minimize variance. In the proposed study, boosted methodology along with Random forest, instead of using only random forest is implemented due to which optimization of the Ant colony search alongside with the proposed classification to hit the classification for sentiment analysis of various reviews of films for research precision.