Improving Calculation of Contextual Similarity for Constructing a Bilingual Dictionary via a Third Language
Takashi Tsunakawa, Yosuke Yamamoto, Hiroyuki Kaji · International Joint Conference on Natural Language Processing · 2013
A novel method is proposed for measuring contextual similarity by “weighted overlapping ratio (WOR)” to construct a bilingual dictionary of a new language pair from two bilingual dictionaries sharing one language. The WOR alleviates the effect of a noisy seed dictionary resulting from merger of two bilingual dictionaries via a third language. Combined use of two word-association measures for extracting contexts from corpora is also proposed to compensate their weaknesses. Experiments on constructing a Japanese-Chinese dictionary via English show that the proposed method outperforms the conventional method based on cosine similarity.