Method for Extracting Unique Meanings of Synonyms Using Word2Vec for Vocabulary Learning

Abstract

In this study, we propose a method for extracting word unique meanings of synonyms using Word2Vec for the purpose of supporting vocabulary learning. In conventional vocabulary learning, it was difficult for the system to teach the meaning differences between synonyms. This study attempts to propose a method for extracting words that represent unique meaning of each synonym pair using Word2Vec. Our method derives a vector of a unique meaning of the target word by subtracting the common vector of the two words from the vector of the target word. In the verification experiment, the proposed method was able to extract unique meaning of Japanese synonym pairs that contain one same Chinese character at a maximum rate of 64%.