Desarrollo de un sistema de recomendación de contenidos educativos basado en estilos de aprendizaje aplicado a la Escuela de Ingeniería en Sistemas.

Abstract

The objective of this work was to develop an educational content recommender system based on learning styles that allows managing and recommending content to students in order to improve their academic performance. Its development contemplated two studies, a preliminary study that established the preferences of content regarding learning styles, and another study to evaluate the academic performance of the students who utilized the system. As data collection techniques, we utilized the following: documentation review, learning style test of David Kolb, test to determine content preference and a structured evaluation with multiple choice questions. The population for the studies were software engineering students from the Escuela Superior Politécnica de Chimborazo, the sample for the first study were 128 students, and for the second study there were two groups of students named control group and experimental group, respectively to apply and compare their results when interacting with the system and its content. We developed the content recommender system with java language programming following the XP methodology and using tools and technologies such as: NetBeans IDE, javascript, Ajax, PrimeFaces and PostgreSQL. The results obtained related to content format preference in relation to learning style demonstrated that the convergent/assimilator groups prefer the video format, while the divergent/accommodator groups prefer the simulation format. The evaluation of academic performance determined that the experimental group obtained an average of 15.60/20, while the control group obtained 12.74/20. We applied the t-student test and we determined that there is a significant difference between the means of the data. We concluded that the use of the recommender system improved student performance by 14.3%. We suggest uploading content in various formats to have more options when making the recommendation.