Learner Corpus-based Research on Automatic Evaluation Methods forReading Ability and Writing Ability

本研究通过收集学习者在写作、发音、阅读理解和听力理解过程中产生的语言数据，构建了一个学习者语料库。这个学习者语料库随后被用来开发对学习者的写作、发音、阅读理解和听力理解的语言技能的自动评估方法。此外，还开发了自动评估方法来衡量每种语言技能的句子难度。对这些自动评价方法的验证实验结果表明，本文提出的学习者语料库和多种语言技能自动评价方法是有效的。

The present study constructed a learner corpus by collecting linguistic data generated in learners' writing, pronunciation, reading comprehension, and listening comprehension. This learner corpus was subsequently used for developing automatic evaluation methods for learners' linguistic skills of writing, pronunciation, reading comprehension and listening comprehension. In addition, automatic evaluationmethods were developed to measure the difficulty of sentences for each linguistic skill. Experimental results for the validation of these automatic evaluation methods showed the efficacy of our learner corpus and automatic evaluation methods for multiple linguistic skills.