Automatic building of lexical resources

OData support
Ács Judit
Department of Automation and Applied Informatics

Nowadays, because of the rapid advance of computers, we use them all around our life. Especially because of the portable and easily accessible devices, it would be important for these devices to be usable in everyday communication.

In most areas of communication, there are well-wersed systems available, but in the field of speech recognition and computational linguistics, most of the technologies are behind the human norms. An average person can easily form sentences, translate from one language to another, and can easily identify the structure of words or a sentence, but for computers, it is not trivial.

A system which could be utilized for computational linguistic purposes is the Wordnet, which conatins meanings, accompanied by words, a definition, and other data that might be valuable. One of these wordnets is the Princeton Wordnet, and the Open Multilingual Wordnet. The main purpose of this document is to examine the methods for automatically expanding these wordnets using dictionaries, and to evaluate the results of these methods. The detailed evaluation will be presented in three languages, however the summary will contain results in over 50 languages.


Please sign in to download the files of this thesis.