1·Large corpora (masses of text) are a good place to start.
通过大型语料库(海量文本)来检查是个好方法。
2·Modeling the linguistic data found in corpora can help us to understand linguistic patterns, and can be used to make predictions about new language data.
建模语料库中的语言数据可以帮助我们理解语言模型，并且可以用于进行关于新语言数据的预测。
3·Supervised classifiers use labeled training corpora to build models that predict the label of an input based on specific features of that input.
监督式分类器使用标签训练语料库来构建模型，预测基于特定要素输入的所输入的标签。
4·Manipulating large corpora, exploring linguistic models, and testing empirical claims.
操作大型语料库，设计语言模型，测试经验假设。
5·As a basis of this study, the present methods are also discussed and summarized from the Angle of corpora in this dissertation.
作为本研究的基础，本文还主要从语料库的角度对现有处理方法进行了讨论和总结。