良質な用例を大規模なコーパスから自動的に抽出できるモデルの構築および試作版の開発

本项目的目的是开发一个模型，从一般语料库中选择教学有效的日语例句，通过调查自动测量的可读性，典型性和信息性的标准。我们从学习词典、参考书、分级读本和学习者语料库中收集例句，构建了一个分级例句语料库，作为验证现有可读性公式在单句或简短用法示例上对日语学习者的可用性的数据集。我们使用现有的可读性公式对这些分级的例句进行了实验，发现虽然这些公式适用于较长的文本，但它们不适用于单个句子。我们进一步注释了一组从网络语料库中提取的句子，手动评分其可读性和信息性，以日语作为外语的学习者，调查可读和信息性句子的可测量标准。对这些标准的分析仍在进行中。我们目前正在探索与构建的单个示例句子和从文本中提取的注释句子集的语料库的可能接口，供日语作为外语的学习者、教师和词典编纂者使用。

The aim of this project was to develop a model for selecting pedagogically valid Japanese example sentences from a general corpus, by investigating automatically measurable criteria of readability, typicality and informativity. We collected example sentences from learners' dictionaries, reference works, graded readers and learner corpora, and constructed a graded corpus of example sentences, to be used as a data set for verifying the usabililty of existing readability formulas on single sentences or short usage examples for learners of Japanese as a foreign language. We experimented using existing readability formulas on these graded example sentences, and found that while the formulas work well for longer texts, they are not applicable to single sentences.We further annotated a set of sentences extracted from a web corpus, manually scoring their readability and informativity for learners of Japanese as a foreign language, to investigate measurable criteria of readable and informative sentences. The analysis of these criteria is still in progress. We are currently exploring possible interfaces to the corpus of constructed single example sentences and the annotated set of sentences extracted from texts to be used by learners, teachers and lexicographers of Japanese as a foreign language.