Combining classification and clustering for tweet sentiment analysis

Abstract

The goal of sentiment analysis is to determine opinions, emotions, and attitudes presented in source material. In tweet sentiment analysis, opinions in messages can be typically categorized as positive or negative. To classify them, researchers have been using traditional classifiers like Naive Bayes, Maximum Entropy, and Support Vector Machines (SVM). In this paper, we show that a SVM classifier combined with a cluster ensemble can offer better classification accuracies than a stand-alone SVM. In our study, we employed an algorithm, named 'C POT.3'E-SL, capable to combine classifier and cluster ensembles. This algorithm can refine tweet classifications from additional information provided by clusterers, assuming that similar instances from the same clusters are more likely to share the same class label. The resulting classifier has shown to be competitive with the best results found so far in the literature, thereby suggesting that the studied approach is promising for tweet sentiment classification.