Integrated Stochastic and Literate Based Driven Approaches in Learning Style Identification for Personalized E-Learning Purpose

Abstract

This paper presents integrated stochastic and literate based driven approaches in learning style identification for personalized e-learning purpose. Shifting a paradigm in education from teacher learning to student learning center has encouraged that learning should follow and tailor learners’ characteristics in the form of personalized e-learning. There are several aspects to describe a condition of learners such as prior knowledge, learning goals, learning styles, cognitive ability, learning interest, and motivation. Even though, in many studies of the personalized e-learning, the learning style plays a significant role. In terms of e-learning, implementing several methods for identifying learner style becomes more challenging. Artificial intelligence and machine learning method give good accuracy, but they still have some issues in computation. Additionally, the stationary method is very hard to represent non-deterministic and dynamic data. Therefore, this research proposes the learning style identification by integrating stochastic and literate based driven approaches. Hidden Markov Model (HMM) and the Naïve Bayes as the Stochastic Approach have been implemented. Subsequently, learner behavior as the literate based data is used to get hints during accessing the learning objects. The proposed model has been implemented to VARK learning style. The accuracy is calculated by comparing the model results with the questionnaire results. When Using the HMM, the proposed model gives accuracy in the range of 95% up to 96.67%. Additionally, when using the Naïve Bayes; the accuracy is 93.33%. The results give better accuracy compared to previous studies. In conclusion, the proposed model is promising for modeling learner style in personalized e-learning.