Aplicação de técnicas supervisionadas e não-supervisionadas da estatística multivariada no contexto da avaliação da efetividade do ensino em um curso de graduação

Abstract

Assuming that the evaluation of learning is not restricted to the quantitative measurement of student learning, it appears his contribution as an analysis tool in various stages of education. Evaluation can increasingly contribute in a meaningful and representative way in student training, where we can highlight some of the factors and be analyzed, such as the influence of the teacher; the school environment; the course curriculum; evaluation mode; among others. Thus, this study presents results of the application of an evaluation instrument of the processes related to student learning, and its consequent application for review and establishment of inferences. The research instrument was based on an existing assessment methodology, called SETE (Students' Evaluation of Teaching Effectiveness), designed to capture the views of students on eleven factors that take into account the effectiveness of teaching. The measuring instrument of this research was applied to students enrolled in undergraduate degree in Computer Engineering from the Federal University of Ceará, and who have completed the first year of the course. Quantified instrument response gave rise to a database, allowing their analysis by two methods of multivariate statistics, Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), nature unsupervised and supervised, respectively. The results showed the feasibility of using the proposed methods to analyze aspects related to evaluation of learning from the perspective of students. This research indicates that the instrument and the proposed methods are able to contribute to the educational evaluation in order to produce results that can generate actions for improvement purposes of the training process of students, serving also as an indicator of the quality of education for the area educational evaluation.