engleski

Automatic Building of a Dictionary of Technical Terms and Collocations Based on AVL Tree

The aim of this work is automatic building of a dictionary of technical collocations. The input to the building procedure is a number of Croatian technical texts from a restricted domain. The dictionary is implemented as an AVL tree, a binary tree which ensures performance of operations such as insertion and retrieval in logarithmic time. Tree nodes contain words and their part of speech (POS) tags. POS tags are partly obtained using the Croatian Morphological Lexicon. POS information enables us to use syntactic filter in order to reduce noise in collocation retrieval.

automatic dictionary generation; AVL tree; collocations

nije evidentirano