This study selects and proposes a list of scoring features to measure formality in the automated grading for KFL speaking. To this end, a total of 462 grammar items were selected by analyzing those presented in Korean grammar dictionaries, research reports, and other relevant materials. Next, the Sejong Corpus was reconstructed to create a formal and an informal reference corpus with 244,344 and 372,585 word segments, respectively. The frequency of occurrence for the analyzed grammar items was calculated, and a statistical significance test (G2) was conducted to assess the differences in frequency in each corpus. The results revealed that 127 grammar items exhibited formality. Considering measurement efficiency, the primary scoring features list was comprised of G2 values of 100 or above, which is the inflection point of G2 value distribution. The list was reviewed and items with inadequate validity were removed. Furthermore, several grammar items were added by referencing formality information from grammar books and G2 values. Through this process, this study constructs and presents a list of scoring features for formality measurement in the automated grading for KFL speaking.
1. 서론
2. 연구 방법
3. 연구 결과
4. 논의
5. 결론
참고문헌
(0)
(0)