نوع مقاله : علمی-پژوهشی
1 دانشگاه علامه طباطبائی
2 کارشناسی ارشد هوش مصنوعی دانشگاه الزهرا
عنوان مقاله [English]
This study explains how to develop the corpus of “ATU Papers, Theses and Dissertations Abstracts” and introduces the different characteristics and features of the corpus. The corpus contains ten thousand thesis abstracts and 9538 article abstracts from the scientific journals of Allameh Tabatabai University with a volume of more than three and a half million tokens. Academic abstracts as brief authored texts with scientific content can depict special linguistic features and therefore, they are valuable documents. In this article, to express the importance of access to such data and to examine some features of the corpus, the word content of a part of the data has been examined and presented according to the concept of keyness and n-grams. The results showed that the lexical content of this corpus could lead researchers to propose some hypotheses. Also, the exploring n-grams of this corpus showed that the language of science has specific word clusters that can depict a particular type of language.