ANALYSIS AND PREDICTION OF THE RELATIONSHIP BETWEEN MULTIPLE INTELLIGENCES IN TWO STEPS: AN INITIAL APPROACH USING MULTIPLE REGRESSION AND MLP

Abstract

The Theory of Multiple Intelligences (TIM) addresses the concept of intelligence, in which their work together promotes the integral development of the individual. Despite not having been developed for educational purposes, TIM started to be widely used in this environment, helping in the teaching and learning process of students. The theory covers eight types of intelligences and has a qualitative character. However, studies appear with the intention of quantifying the approach of the Intelligences. Based on this approach, this article aims to quantitatively relate the intelligences to each other and verify how a prediction model reacts to this data set. For this, multiple regression and an artificial neural network (Multilayer Perceptron — MLP) were used. This research, of an applied nature, was carried out with students from the three grades of high school in a state public school in Fortaleza/CE. They answered a questionnaire with 80 questions based on the Multiple Intelligences itinerary developed by Armstrong (2001). A set of eight multiple regressions was created and the presented results show that there were small distortions in some coefficients when compared with the indices of the correlations between the variables. One of the plausible hypotheses is the size of the dataset. On the other hand, the prediction achieved with the MLP proved to be promising in terms of analyzing the development situation of any student. It is intended, in future studies, that the results can be implemented in a chatbot that offers, in a predictive way, prescriptions for the development of intelligences in the school environment.