Clustering Opinion Polarity of Student’s Strategies as Clues for Motivation

Nowadays education is considered critical and indispensable for the country’s development. Understand student’s motivation for study is one of the most challenging task in educational process due to the high level of diversity present in the students and educational environment. For this reason, in this paper we propose the use of patterns recognition to solve this problem. A combination of a PCA + K-means algorithm was presented and a dataset of different opinion for students strategies in lectures (Challenging, Non-Challenging and Generic) was used. As a result, reaching optimum values for different clustering metrics and including a Silhouette index over 0.55 was possible to create clusters that explain the existing patterns in the student’s opinion for different study strategies in lectures. Also, an explainable tool (SHAP) was used to extract knowledge from the clusters created which suggested clues related to students.