Text Sentiment Classification Based on Mixed Cloud Vector Model Clustering and Kernel Fisher Discriminant

In today’s world, the web has dramatically changed the way that people express their opinions. People use the internet to express their opinion, attitude, feeling and emotion about films, goods, news etc. It is challenging to automatically classify mass subjectivity comments into different sentiment orientation categories (e.g. positive/negative). Furthermore, the ambiguity and randomness, which are existed in natural language, lead to lower classification accuracy in text sentiment classification. In this paper, we propose a novel chinese text sentiment classification algorithm based on mixed cloud vector model clustering and kernel fisher discriminant. In this algorithm, we firstly analysis the role of cloud model theory in conversion between qualitative concept and quantitative values, and explore a mixed feature cloud model (MFCM) based on cloud model to represent a single document. In MFCM, both effect of different part-ofspeech features and ambiguity of sentiment tendency are considered. And then, documents are clustered according to their similarity between MFCM. Finally, kernel fisher discriminant (KFD) is adopted as the classifier to judge views. The experimental results demonstrate that our proposed method outperforms traditional approaches. Keyword: Text sentiment classification; cloud model; kernel fisher discriminant; support vector machine