Using Appraisal Taxonomies for Sentiment Analysis

Recent years have seen a growing interest in non-topical text analysis, in which characteri- zations are sought of the opinions, feelings, and attitudes expressed in a text, rather than just the facts. A key problem in this area is sentiment classification, in which a document is labelled asapositive('thumbsup')ornegative('thumbs down') evaluation of a target object (film, book, product, etc.). Immediate applications include data and web mining, market research, and cus- tomer relationship management.