LewiSpace: an Exploratory Study with a Machine Learning Model in an Educational Game

Abstract

The use of educational games as a tool for providing learners with a playful and educational aspect is widespread. In this paper, we present an educational game that we developed to teach a chemistry lesson, namely drawing a Lewis diagram. Our game is a 3D environment known as LewiSpace and aims at balancing between playful and educational contents in order to increase engagement and motivation while learning. The game contains mainly five different missions aim at constructing Lewis diagram molecules which are organized in an ascending order of difficulty. We also conducted an experiment to gather data about learners' cognitive and emotional states as well as their behaviours through our game by using three types of sensors (electroencephalography, eye tracking, and facial expression recognition with an optical camera) and a self report personality questionnaire (the Big Five). Primary results show that a machine learning model namely logistic regression, can predict with some success whether the learner will success or fail in each mission of our game, and paves the way for an adaptive version of the game. This latter will challenge or assist learners based on some features extracted from our data. Feature extraction integrated into a machine learning model aims mainly at providing learners' with a real-time adaptation according to their performance and skills while progressing in our game.

From substitution to redefinition: A framework of machine learning‐based science assessment

“…With timely feedback, the game could be adapted according to students' performance on‐the‐fly to accommodate their abilities. Ghali et al (2016) reported that using data collected from affective electroencephalography, eye tracking, facial expression, and self‐report, they could predict students' problem‐solving behavior using ML. They also explored the possibility of using ML to adapt to users' abilities in real time.…”

“…These programs suggest that using technology to elicit performance on scientific inquiry practices—such as identifying principles, using principles, and conducting inquiry—has significant advantages over traditional paper and pencil tests (Quellmalz et al, 2013). Technology‐based assessments also permit harvesting information associated with assessment tasks (e.g., mobile eye‐tracking systems, emotional expression systems, and brainwave systems (Chanijani et al, 2016; Ghali, Ouellet, & Frasson, 2016). For example, employing the NeuroSky brainwave system, Chen, Wang, and Yu (2017) used sensors to detect students' brainwave signals and accurately predict attention levels during learning.…”

Multi-Level Adaptation of an Educational Game to Individual Student’s Gameplay, Knowledge and Emotions

“…In many cases, knowledge assessment can lead to negative emotions, e.g., anxiety or fear, if a learner is not confident about his/her knowledge level or if the selected challenge level is not matched appropriately to learner's skills. Analysis of the latest research reveals that despite numerous studies relating emotions with games [3,4], only in rare cases emotions are actually considered for the adaptation of game elements [5,6]. Furthermore, an adaptation itself in games is a rare phenomenon.…”

“…Games are considered to be beneficial for learning because they incorporate two fundamental aspects: (1) educational aspect related to the learning content and strategies presented to learners, and (2) playful aspect that allows players to act, explore, take rewards, etc. [3]. Games are internally motivating, involve active cognitive, physical, and affective engagement allowing players freely to experiment, fail and recover from failures by trying over and over again [7].…”

The Adaptation Approach for Affective Game-Based Assessment

“…The analysis of the latest research shows that regardless of numerous studies that link together emotions and games [22], [23] only in rare cases emotions are analysed for the adaptation purposes. The mathematics game called Heroes of Math Island includes an affective agent (personified by a monkey character) that uses emotional expressions to respond to situations in the game [24].…”