Motivation and personalization of teaching with machine learning

Abstract

The motivation of the student causes the teaching experience to be more enjoyable for the student and results in better utilization of the teaching activity. The key is to identify where that motivation lies in order to adapt the content to the student's expectations. The objective of this work is to establish a method to identify the student's motivation regarding the training they are going to receive and be able to personalize the learning experience according to this motivation. To achieve this, we describe an experience in which a machine learning model of decision trees was trained using a voluntary survey generated through LinkedIn. By consulting the LinkedIn profiles of the respondents, a training dataset was created, which resulted in a model that achieved a 72% accuracy rate in a 10-fold stratified cross-validation. During the presentation of the students who enrolled in the activity, the necessary information was captured to generate a test dataset, which was used to validate the trained model. The accuracy rate of this validation was 100%. Although the sample size and predictors used are limited, we believe that this experience sufficiently illustrates the potential of artificial intelligence to identify student motivations and thus personalize the teaching experience, with the aim of increasing motivation and improving student performance.