Research On Modeling Gamified Learning Behaviors Based On Hidden Markov Models

During the post-pandemic era,colleges and universities have launched more online courses,such as MOOCs or SPOCs,than ever because teachers and students demand distant education.Learning management systems for such a circumstance become more urgent.However,online learning requires learners’ activeness.In other words,low learning motivation may result in low participation and ineffective learning.In recent years,gamified learning has attracted the attention of many researchers due to its positive role in promoting learners’ motivation,learning participation,collaboration,and effectiveness.By adding gamification mechanisms to learning,we can create a sense of immersion similar to games and improve learners’ learning experience,which involves positive learning behavior.Learning behavior modeling as an emerging method is conducive to broadening our understanding of learners’ learning behavior patterns in a gamified learning environment.In this thesis,the author uses a self-developed online academic reading and writing system as a research tool with gamified mechanisms.Besides,the author uses hidden Markov models as an analysis tool for modeling gamified learning behaviors.More specifically,the main research work is described as follows:First,the literature related to gamified learning and learning behavior modeling technology are reviewed to justify the reasons for choosing Hidden Markov Model as the analysis tool in this research.Then,the three basic problems and related issues in Hidden Markov Models are discussed.After the corresponding solution algorithm is explained in detail,the scale factor is added and the revaluation formula is revised for the actual problems faced in this research.The optimal number of hidden states of the model is selected according to the maximal likelihood,and combined in the model parameter estimation.K-means algorithm optimization is also performed for model optimization to improve the slow convergence speed of the Baum-Welch algorithm and the ease of reaching a local optimal solution in the modeling process.Secondly,an online academic reading and writing system was developed to support the teaching of postgraduate literature reading and writing courses.Besides,gamified learning theory is also applied to design three gamification mechanisms,i.e.badges,points,and rankings,and to integrate them with the system.For developing the online academic reading and writing system,the author chooses the front-end and back-end separation method.Furthermore,the React framework is adopted to complete the development of the front-end and the ThinkPHP framework is adopted for the back-end.More specifically,there are five functional modules in the system:literature search,literature reading,paper writing,classroom participation,and peer review.While these modules are designed to assist learners to complete paper writing,the gamification mechanisms in the system are provided to enhance learning motivation and the sense of participation.Finally,with 94 first-year graduate students as subjects,a total of 21,663 online learning behavior data were collected in the online academic reading and writing system.Based on hidden Markov models,the learning behavior models of learners in a non-gamified and gamified learning environment were constructed.The results of comparative analysis show that students in the gamified learning environment seem more motivated to learn,and have more social interactions with their peers in terms of learning strategies.More importantly,they can revise their papers with the help of their peers’ opinions and self-reflection.