Sentimental analysis of tweets using Ant Colony Optimizations

Sentiment analysis is a form of information extraction from text of growing research and commercial interest. In this paper, we present our machine learning experiments in regarding to sentiment analysis on public tweets, which is collect by tweet programmer account then extracting the features after the preprocessing of text then given the relative weight to every features. These features are learned by two types of classifiers: one is generative classifier (Naive Bayes) which understands why some text is positive sentiment and others are negative. Second classifier is discriminative classifier (Support vector machine (SVM)). In proposed methodology, features scarcity reduced by optimization which reduces the complexity of processing and learns features according to its relative weight leads to the reduction of overlapping in features. In this paper, our experiment clearly shows the difference between optimized and without optimized features in case of SVM but to some extent in Naive Bayes.