Sentiment Augmented Bayesian Network

Orimaye, S.O.

Sentiment Classification has recently gained attention in the literature with different machine learning techniques performing moderately. However, the challenges that sentiment classification constitutes require a more effective approach for better results. In this study, we propose a logical approach that augments the popular Bayesian Network for a more effective sentiment classification task. We emphasize on creating dependency networks with quality variables by using a sentiment-dependent scoring technique that penalizes the existing Bayesian Network scoring functions such as K2, BDeu, Entropy and MDL. The outcome of this technique is called Sentiment Augmented Bayesian Network. Empirical results on three product review datasets from different domains, suggest that a sentiment-augmented scoring mechanism for Bayesian Network classifier, has comparable performance, and in some cases outperform state-of-the-art sentiment classifiers.

Cite as: Orimaye, S.O. (2013). Sentiment Augmented Bayesian Network. In Proc. Eleventh Australasian Data Mining Conference (AusDM13) Canberra, Australia. CRPIT, 146. Christen, P., Kennedy, P., Liu, L., Ong, K.L., Stranieri, A. and Zhao, Y. Eds., ACS. 89-97

(from crpit.com) (local if available)