Taxonomy Learning: a Survey of Approaches
Elly van den Heuvel · 2009
The last few years the field of taxonomy learning has become more and more important in computer science. However it also has relevance in economics, as it can be used for describing economic domains as finances, macroeconomics etc. There are numerous methods to create a taxonomy from the Web, corpus, or semantic lexicon. This thesis examines the different steps of the investigated methods dealing with the process of taxonomy learning. These methods will be explained and compared throughout this thesis. The main methods that are discussed are term extraction through linguistic, statistic and hybrid filtering, synonym extraction, computing distance between terms, and deriving taxonomic relations through clustering, classification and lexico-syntactic patterns.