Descriptions of the DocuScope tags for the different versions of the dictionary.
DocuScope
DocuScope is a dictionary-based tagger, developed by David Kaufer and Suguru Ishizaki at Carnegie Mellon University.\n It consists of an enormous lexicon organized into a 3-level taxonomy. An analogue would be the lexicons typically used in sentiment analysis. Those usually organize words and phrases into 2 categories (positive and negative) and work by matching strings over a corpus of texts.\n DocuScope works in the same basic way, but organizes its strings into many more categories and is orders of magnitude larger. A typical sentiment lexicon may match 3-5 thousand strings. DocuScope matches 100s of millions. You can find a small, early version of the dictionary here.