标题：Sentence-Level Sentiment Analysis in the Presence of Modalities
作者：Liu, Yang; Yu, Xiaohui; Liu, Bing; Chen, Zhongshuai
作者机构：[Liu, Yang; Yu, Xiaohui; Chen, Zhongshuai] Shandong Univ, Sch Comp Sci & Technol, Jinan 250100, Peoples R China.; [Liu, Bing] Univ Illinois, Dept Co 更多
会议名称：15th Annual Conference on Intelligent Text Processing and Computational Linguistics (CICLing)
会议日期：APR 06-12, 2014
来源：COMPUTATIONAL LINGUISTICS AND INTELLIGENT TEXT PROCESSING, CICLING 2014, PART II
关键词：Opinion mining; Sentiment analysis
摘要：This paper studies sentiment analysis of sentences with modality. The aim is to determine whether opinions expressed in sentences with modality are positive, negative or neutral. Modality is commonly used in text. In a typical corpus, there are around 18% of sentences with modality. Due to modality's special characteristics, the sentiment it bears may be hard to determine. For example, in the sentence, this cellphone would be perfect if it has a bigger screen, the speaker is negative about this phone although there is a typically positive opinion word "perfect" in this sentence. This paper first presents a linguistic analysis of modality, and then identifies some features to train a support vector machine classifier to determine the sentiment orientation in such sentences. Experimental results on sentences with modality extracted from the reviews of four different products are given to illustrate the effectiveness of the proposed method.