Good Examples for Terminology Databases in Translation Industry

This paper deals with finding good examples for terminology database entries in the translation industry. When extracting terms from bilingual translation memory exchange files, it is very easy to also extract example sentences to showcase the use of the term in practice. However, there are usually a lot of sentences containing the term and selecting an appropriate example is not a straightforward task. In this paper, we explore the use of data mining techniques to find good term examples. After constructing the corpus from a large English-Slovenian bilingual file from a financial domain, we extract linguistic features and load them into the Weka data mining environment to analyze the performance of various classifiers, resulting in 0.8 precision for positive class (good examples) and 0.85 overall accuracy. While the model was tested only on one language combination, the nature of most features is language-independent which suggests that the model could be used successfully for other language combinations