Self-reflection is an effective way in teachers' professional development. How to evaluate the content of self-reflection automatically is a critical issue in the web-based software system. The content of self-reflection submitted under web environment is pre-processed such as Chinese word segmentation, stop words filtering and so on, build self-reflection text vector with vector space model, calculate the similarity between self-reflection text and corpus text based on cosine theory. According to the level of corpus text with maximum similarity and previous threshold of system,achieve the automatic evaluation of self-reflection content. The experiment shows that the correct rate of the automatic evaluation result is more than 90% compared with the expert evaluation result,implement the automatic assessment of self-reflection content well.%教学反思是教师专业能力发展的重要途径,对反思内容进行自动评估是网络环境下教学反思系统亟待解决的关键问题.对网络环境下提交的反思文本进行中文分词、停用词过滤等预处理,采用向量空间模型构建反思文本向量,基于余弦理论计算反思文本与语料库文本的相似度.根据最大相似度语料文本的等级及系统预设阈值,实现反思内容的自动评估.实验结果表明,自动评估结果和专家认定的评估结果相比,正确率达到90％以上,基本实现了反思内容的自动评估.
展开▼
