A CHINESE VOCABULARY LEARNING SYSTEM: LATENT SEMANTIC ANALYSIS APPROACH

摘要

The present study aims on developing an LSA-based Chinese Vocabulary Learning System to learn Chinese words effectively and adaptively. Chinese words that were introduced in lessons were selected systematically by the statistical computation method of LSA (Latent Semantic Analysis). By the means of LSA, a list of prioritized vocabulary for different age groups will be provided by computing the relevancy between words in texts and Chinese children corpus. A total of 51 second graders (23 in the control group, 28 in the experimental group) underwent a γ-week intervention by two different vocabulary learning methods, traditional vocabulary learning and LSA-based vocabulary learning. Peabody Vocabulary, Chinese Character Recognition, and Word Association tests were administered before and after the intervention. The results of paired t-tests showed that the experimental group improved significantly after the 7-week intervention among tests, whereas none of the test improved significantly in the control group. Moreover, the results of ANCOVA indicate that after controlling for the effects of pretests, the experimental group performed significantly better than the control group on Chinese Character Recognition and Word Association. The results suggest that LSA-based vocabulary learning system is effective for vocabulary development among children who are learning to read Chinese.