ЛИНГВИСТИЧНО ОПИСАНИЕ И КОМПЮТЪРНА ОБРАБОТКА НА СЪСТАВНИ ИМЕНУВАНИ СЪЩНОСТИ В БЪЛГАРСКИЯ ЕЗИК

Abstract

We present an automatic approach for the creation of a Bulgarian Morphological Dictionary of Named Entities consisting of 21,161 named entities (NEs). Each NE is assigned an inflectional type which reflects its part-of-speech, lexico-grammatical and morphological properties, including the morphonemic alternations. NEs are classified into semantic categories according to an extended hierarchical system. The wordforms of each entry are generated automatically from its inflectional type, and the relevant grammatical characteristics are assigned. The result is validated by experts.