Presenting a Novel Hybrid Approach of Text Mining Sentiment Analysis in Twitter Using CART Decision Tree

Today, with the enormous growth of the Internet and social networks as virtual communities and mass media, and increased use of them, a huge amount of user feedback comes from a variety of topics. Therefore, the use of novel approaches for analyzing them seems to be necessary. Text mining, as a special strategy, drives the knowledge discovery process, which uses non-verbal and attractive patterns of natural language processing. In this paper, a new hybrid approach of machine learning and vocabulary-based method to text-mining sentiment analysis on Twitter. To improve text-mining and sentiment analysis, the CART decision tree is used as a machine learning method for classification, also for extracting more precisely sentiment, we use from the list of SentiStrength algorithms as a lexicon-based method. CART is very effective in processing discrete and continuous data in text mining. The unique CART feature is a complex data structure analysis that can support regression as well as classification operations, according to the input of the problem. The ability and power of the SentiStrength algorithm to detect sentiment has also led to a thorough analysis of sentiment in tweets. The results of the implementation in the polarity recognition show improvement of classification in the most feature.