摘要
国际中文教育的文本可读性研究基于对汉语文本难度进行定量分析,以直观反映语料的难易度,主要采用公式法和分类模型法。构建可读性公式是通过选取与可读性密切相关的文本特征,采用多元线性回归的方法,来建立特征因素与可读性级别之间的函数关系;基于分类的研究,则采用支持向量机和朴素贝叶斯等方法进行模型构建。文本特征的选用和测定是可读性研究的重点,字、词、句层面的文本特征是影响可读性的重要因素。随着研究方法的更新、研究内容的深入,必将会出现评估结果更为精确、使用范围更为广泛、也更适合汉语文本的新的自动评估方法。
The study of text readability in International Chinese Language Education quantitatively analyzes the difficulty of Chinese texts and visually reflects the difficulty level of the corpus.Related studies mainly use formula and classification model methods.The readability formula is constructed by selecting text features closely related to readability and using multiple linear regression methods to establish a functional relationship between feature factors and readability levels,and classification-based studies use methods such as support vector machines and plain Bayesian methods for model construction.The selection and measurement of text features are the focus of readability studies,and text features at the word,phrase and sentence levels are important factors affecting readability.The results can inform the creation of text readability studies that are more accurately assessed and more widely used.
作者
张庆翔
张莹
Zhang Qingxiang;Zhang Ying(College of Liberal Arts,Shanghai University,Shanghai 200436,China)
出处
《现代语文》
2022年第10期89-95,共7页
Modern Chinese
关键词
文本可读性
可读性公式
可读性模型
国际中文教育
text readability
readability formula
readability model
International Chinese Language Education
