Developing and Evaluating a Computer-Assisted Near-Synonym Learning System Using Multiple Contextual Knowledge Sources

ABSTRACT

Despite their similar meanings, near-synonyms may have different usages in different contexts. For second language learners, such differences are not easily grasped in practical use. In this paper, we develop a computer-assisted near-synonym learning system for Chinese English-as-a-Second-Language (ESL) learners to better understand different usages of various English near-synonyms in a range of contexts. To achieve this goal, we implement the system using two automatic near-synonym choice techniques: pointwise mutual information (PMI) and n-grams to provide useful contextual information for learning frequent and discriminative words and n-grams occurring in the context of different near-synonyms. The system is evaluated using a vocabulary test with near-synonyms as candidate choices. Participants are required to select the best near-synonym for each question both with and without use of the system. Experimental results show that both techniques can improve participants’ ability to distinguish different usages of various near-synonyms in a range of contexts, and use them appropriately. In addition, participants are found to prefer to use the PMI in the test, despite n-grams providing more precise information.