The purpose of this study is to evaluate the level of technical completeness of classifying text by subject through analysis of Korean natural language processing(NLP) open-source software related to the currently developed text classification. To this end, after searching for the released NLP software, the accuracy was evaluated using Naver news articles. As a result, it was analyzed that Simple RNN had an accuracy of 78% and LSTM of 80%. In addition, the model was evaluated using <Cyber Korean Intermediate 1&2>, and it was difficult to determine the accuracy because the training data was different from the evaluation data. However, if text classification technology is applied to Korean language education based on this study, the significance of this study can be found in that it is possible to efficiently manage vast amounts of Korean reading materials and provide texts tailored to learners’ interests.
1. 서론
2. 텍스트 분류 인공지능 기술
3. 연구 내용 및 방법
4. 분석 결과
5. 결론
참고문헌
(0)
(0)