Abstract:

Based on the analysis on the linguistic features of Turkish military terminologies, we proposed a terminology extraction method using combination of rules and statistics algorithm, and experimentally verified this method on monolingual W-data and N-data. The candidate items in the monolingual text are filtered through keywords, stop words, morphological analysis sequence pattern, pointwise mutual information, left and right information entropy and adjacency suffixes. Our results show that the proposed method can effectively extract Turkish military terminologies from the experimental data.

Key words: Turkish terminology extraction, filtering dictionary, automatic extraction methodologies