Abstract:

Based on the attitudinal system of appraisal theory and the semantic annotation method from corpus study, this paper aims to discover the parameter features of Chinese online commentaries in building up their sentiment meanings. It is found that multi sentiment parameters work together during the sentiment expression of Chinese online commentaries. This annotated corpus will not only provide rich structured semantic sentiment resources for Chinese sentiment computation to facilitate machine learning, opinion extraction and summarization, but also offer a new multi-parameter operation system for present Chinese sentiment annotation work in computer science and linguistics domains.

Key words: sentiment corpus, Chinese, sentiment parameter, appraisal theory, online commentary