National Engineering Research Center for E-Learning, Huazhong Normal University, Wuhan, China

---

Abstract—Online review websites provide an important channel for people to share their opinions. In this paper, we research the sentiment recognition technology on online Chinese micro movie reviews. As the sentiment expressed in Chinese is subtle and the feature space is very sparse, we adopt n-grams to develop sentiment feature space, and then propose an ensemble learning algorithm based on random feature space division method, namely Multiple Probabilistic Reasoning Model (M-PRM), for supervised document level sentiment classification. This algorithm captures discriminative sentiment features and makes full use of them. Comparing with this algorithm, we apply other four machine learning methods: Multinomial NaiveBayes (MNB), Probabilistic Reasoning Model (PRM), Sentimentword method (SWM) and SVM on two micro movie review datasets. Results show that M-PRM achieves better classification performance than other methods.

Index Terms—sentiment recognition, Chinese movie review, M-PRM, ensemble learning

[PDF]

Cite: Wei Xu, Zhi Liu, Tai Wang, and Sanya Liu, " Sentiment Recognition of Online Chinese Micro Movie Reviews Using Multiple Probabilistic Reasoning Model," Journal of Computers vol. 8, no. 8, pp. 1906-1911, 2013.