A model of textual affect sensing using real-world knowledge

摘要: This paper presents a novel way for assessing the affective qualities of natural language and scenario its use. Previous approaches to textual affect sensing have employed keyword spotting, lexical affinity, statistical methods, hand-crafted models. demonstrates new approach, using large-scale real-world knowledge about inherent nature everyday situations (such as "getting into car accident") classify sentences "basic" emotion categories. commonsense approach has robustness implications.Open Mind Commonsense was used real world corpus 400,000 facts world. Four linguistic models are combined society commonsense-based recognition. These cooperate compete text. Such system that analyzes sentence by is practical value when people want evaluate text they writing. As such, tested in an email writing application. The results suggest robust enough enable plausible user interfaces.