Sentiment classification by a hybrid method of greedy search and multinomial naïve bayes algorithm

In this paper, we proposed sentiment classification framework focusing on the hybrid method of greedy and multinomial naive bayes algorithm. We found greedy search feature selection most effective in our experiments with multinomial naive bayes algorithm. We also discovered that the multinomial naive bayes is suitable for combination with the greedy method. The hybrid method of greedy and multinomial naive bayes algorithm yielded the best performance with the accuracy over all traditional algorithms. Based on our experiments, the multinomial naive bayes algorithm with the greedy search feature selection yielded the best performance with the accuracy of 85.00 %. Our experimental results also reveal that hybrid methods have a positive effect on sentiment classification framework.