Texts and emoticons expressing sentiment can be used to analyse emotion. In an Internet environment, emoticons are frequently used, which have explicated information for emotion analysis. Considering the characteristics of short texts including sparseness, non-standardization and ambiguities in a subject, two models based on word embedding, emotion-dictionary and fuzzy reasoning are proposed: the low-dimensional hybrid feature model and the emotion-enhanced inference model. The low-dimensional hybrid feature model includes the number of emoticons, the emotion-word number and the negative-word number in a text. The emotion-enhanced reference model includes some fuzzy reasoning rules and a variety of the combinations of emotion-words, negative-words, and question marks and exclamation points. The validity of the model has been verified based on Douyin reviews and the data of the 2nd CCF Conference on Natural Language Processing and Chinese Computing (NLPCC 2013), where the average accuracy rate on Douyin reviews achieved is \(89.16\%\). Through the comparative experiment, the results show that the models are more effective in ultra-short emotion text classification than the comparison models.

表达情感的文本和表情符号可用于分析情感。在互联网环境中，表情符号被频繁使用，它具有用于情感分析的解释信息。针对主题中短文本的稀疏性、非标准化和歧义性等特点，提出了两种基于词嵌入、情感字典和模糊推理的模型：低维混合特征模型和情感增强推理模型。低维混合特征模型包括文本中的表情符号数、情感词数和负词数。情感增强参考模型包括一些模糊推理规则和情感词、否定词、问号和感叹号的多种组合。\(89.16\%\)。通过对比实验，结果表明，该模型在超短情感文本分类方面比对比模型更有效。