SOFTWARE FRAMEWORK FOR NATURAL LANGUAGE PROCESSING

V magistrskem delu predstavimo področje obdelave naravnega jezika in besedilnega rudarjenja ter opišemo opravila obeh področij. Nadaljujemo z opisom generičnega programskega paketa GATE za procesiranje naravnega jezika, predvsem besedil v angleškem jeziku. V glavnem delu predstavimo lastno implementacijo razširljivega programskega paketa za procesiranje besedil v slovenskem jeziku, kjer predstavimo interno zgradbo programskega paketa, podamo podroben primer implementacije programskega dodatka in zgradbo dokumenta, ki hrani rezultate procesiranja. Dodatno opišemo pomembnejše programske dodatke, ki smo jih izdelali. Delovanje programskega paketa predstavimo na konkretnih primerih uporabe.Research field of natural language processing and text mining is presented along with some of the tasks performed in both fields of research. We continue with description of a generic natural language processing software package named GATE, primarily developed for English language. In the main part of the document we present our own implementation of extensible software package for natural language processing, primarily for Slovenian language. Internal structure of the software package is presented, along with detailed example of software plug-ins implementation and resulting document structure, which carries all the results produced by any plugin. Additionally, some important plugins are described. Concrete usage scenarios of the software package are also presented