BUILDING A CORPUS BASED ADJECTIVE LEXICON FOR TURKISH

Abstract

This paper describes the design and construction of a lexical database for Turkish adjectives. We used a textual corpus of about one million running words that we collected from on-line newspapers and magazines available on the Internet. The lexicon contains syntactic category, semantic category, gradability, and thesaurus information about adjectives as well as selectional restrictions. It supports Natural Language Processing (NLP) applications such as parsing, text generation, natural language understanding, and information retrieval. It has been implemented as a relational database. The process of building the lexicon from the textual corpus has been performed semi-automatically using a series of extraction programs. We also implemented a Graphical User Interface to the lexicon.