A text as unique as fingerprint
Marcella Rocha, Giovani Ângelo S. da Nóbrega, Ricardo Alexsandro de Medeiros Valentim, Luca Pareja Credidio Freire Alves · 2020
Authorship Attribution, study to deduce an author for a given text based on then writing characteristics is a problem with a long history and refers to the task of author recognition of a text in a group of candidate authors, based on relevant features extracted from the text (Stylometry). Stylometry occurs by identifying relevant attributes that define a space in which authors can be distinguished from each other. Writers use language in different ways to express their ideas, and this variation in language makes authorship recognition possible. In this paper the problem of authorship recognition is being exposed in order to make it a tool for use in the distance education platform of the Ministry of Health: AVASUS. Therefore, the stylometric characteristics are extracted from the collected data set and different classification algorithms are trained to predict the authors of texts longer than 30 characters. The results were relevant obtaining values of 92% accuracy in some classifiers and showed that extraction techniques of stylometric characteristics could be used to recognize the author of a text.