In this paper, we reproduce some of the experiments related to bilingual terminology alignment described by Aker et al. (2013). They treat bilingual term alignment as a binary classification problem and train a SVM classifier on various dictionary and cognate-based features. Despite closely following the original paper with only minor deviations - in areas where the original description is not clear enough - we obtained significantly worse results than the authors of the original paper. In the second part of the paper, we try to analyze the reasons for the discrepancy and offer some methods to improve the results. After improvements we manage to achieve a precision of almost 91% and recall of almost 52% which is close to the results published in the original paper. Finally, we also performed manual evaluation where we achieved results similar to the original paper.
@InProceedings{REPAR18.1, author = {Andraz Repar ,Matej Martinc and Senja Pollak}, title = {Machine Learning Approach to Bilingual Terminology Alignment: Reimplementation and Adaptation}, booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year = {2018}, month = {may}, date = {7-12}, location = {Miyazaki, Japan}, editor = {António Branco and Nicoletta Calzolari and Khalid Choukri}, publisher = {European Language Resources Association (ELRA)}, address = {Paris, France}, isbn = {979-10-95546-21-4}, language = {english} }
Summary of the paper