The objective of this PhD project was to propose the design of an online corpus-driven dictionary of Portuguese for university students (DOPU), aimed at both speakers of Portuguese as a mother tongue and as an additional language and covering Brazilian and European Portuguese varieties. For that, the highly innovative semi-automated approach to dictionary-making (Gantar, Kosem and Krek 2016) was adopted, which involves automatic extraction of data from the corpus and import into dictionary writing system. As a method that had never been applied for lexicographical projects of the Portuguese language, it was necessary to experiment the approach for the first time. Thus, all the required pre-requisites were newly developed, namely, a corpus of academic texts, sketch grammar, GDEX configuration, and a specially-tailored procedure for automatic extraction of data. The experiment indicated that not only can this approach be successfully used as a means to provide lexical content for the design of DOPU, but it can also be beneficial to other lexicographical projects of Portuguese.

中文翻译：

---

面向学生的葡萄牙语在线语料库驱动词典的设计建议

该博士项目的目的是为大学生提出在线葡萄牙语语料库驱动的葡萄牙语词典（DOPU）的设计，该词典针对既讲葡萄牙语又作为母语和另一种语言的葡萄牙语使用者，涵盖巴西和欧洲葡萄牙语品种。为此，采用了高度创新的半自动化字典制作方法（Gantar，Kosem和Krek，2016年），该方法涉及从语料库中自动提取数据并将其导入字典编写系统。作为从未在葡萄牙语的词典项目中应用的方法，有必要首次尝试这种方法。因此，所有必要的先决条件都是新开发的，即学术文献集，草图语法，GDEX配置，以及专门为自动提取数据而定制的过程。实验表明，这种方法不仅可以成功地用作为DOPU设计提供词汇内容的手段，而且对葡萄牙语的其他词典项目也有好处。