Modeling Sentiment Analysis for Educational Texts by Combining BERT and FastText
Pan Xie, Hengnian Gu, Dongdai Zhou · 2024
With the development of artificial intelligence technology, the field of education has begun to apply text mining technology to education management, teaching evaluation, and learning support. The analysis of the text generated during the learning process by educational text mining helps to better grasp the students' internal psychological characteristics such as emotions, thinking, and cognition, providing evidence support for precise teaching design. Comparing various methods of educational text sentiment analysis, we propose an educational text sentiment analysis model that combines BERT and FastText, which effectively solves problems such as polysemy and new words on the internet. Taking the teaching text data of ideological and political courses on online education platforms as the research object, the educational text sentiment analysis model combining BERT and FastText will provide more precise and efficient solutions for analyzing the dynamic changes in students' online ideological and political learning emotions.