Semantic Tagging at the Sense Level

This paper summarizes our research in the area of semantic tagging at the word and sense levels and sets the ground for a new approach to text-level sentiment annotation using a combination of machine learning and linguisticallymotivated techniques. We describe a system for sentiment tagging of words and senses based on WordNet glosses and advance the treatment of sentiment as a fuzzy category.

• Type: Conference
• Year: 2006
• Where: AAAI
• Authors: Alina Andreevskaia, Sabine Bergler
• Keywords: AAAI 2006, Intelligent Agents, Sense Levels, Text-level Sentiment Annotation, WordNet Glosses