Mineração de poemas através de técnicas de processamento de linguagem natural.

Abstract

The contact with poems in basic education is an encouragement for students to discover the pleasure provided by the experience with poetic language. The Coletânea de Poesias project, held annually at Fera Colégio e Curso, is an initiative that aims to promote this contact through the reading, appreciation, and writing of poems, generating yearly a book with texts written by students of the secondary education. Given the manual analysis of these texts would be costly, the present work employed techniques of Natural Language Processing, such as Part-of-Speech tagging and topic modeling, to analyze the poems produced in the ten most recent editions of the project. The results obtained reinforce aspects related to the freedom of creation involved in poetic production and that the themes addressed by the students vary according to their maturity and their environment.