Comparison of selected methods for the retrieval of neologisms

The paper discusses and compares several semi-automatic methods used to extract neologisms from linguistic corpora. All the methods are based on the concept of discriminants, or textual features (both lexis and punctuation), that either precede (lexical discriminants) or confine (punctuation discriminants) phrases in which the occurrence of neologisms is higher than elsewhere in the text. Excerption and comparison was conducted on a corpus of 45 million words, articles from Nature scientific magazine. The putative neologisms were extracted using morphological analysis and frequency of their occurrence in the Google search engine. The result is a list of 1000 neologisms and assessment of the efficacy of each method.