Research On Learner Model For Personalized Education

Nowadays, with the rapid developments of Computer technology and Internet, education has been shifting from the traditional mode to the web-based Distance Education mode which introduces the multimedia technology. Distance Education breaks the time limit and the space constraint to traditional education. Due to its quick and timely information exchange capability and wealth of knowledge forms, Distance education has become an important way to acquire knowledge and enhance ability for people in the modern life, and therefore earned more and more attention.Currently, research on Distance Education system are focus on intelligentand personalization, and a key issue is how modeling learners in education system. If learner model can accurately and timely reflect the actual situation of the learner?s study is the base for arranging learning plan, selecting teaching strategies and pushing learning resources in the system.This paper focus on Learner Modeling Technology for Personalized Education, one contribution of this thesis is proposing a Learner Model-LMFPE, which contains three sub-models named Learning Interest Model-SIMFPE, Knowledge Model–KMFPE and Cognitive Mode-CoMFPE. Combined with previous research, not only all the three sub-models? modeling techniques are discussed, but also the status and role of these models in Personalized Education are analyzed.Meanwhile, for Learner Knowledge Model, which Reflects the learning progress and knowledge mastery level of learners, we propose KMFPE, which constructed by Bayesian network and whose nodes are knowledge items and question items from education. KMFPE supports using learner?s testing for model?s online evolution, which can maintain the system?s accrual description of learner?s knowledge mastery level. Online evolution algorithm is an application of Bayesian network?s online learning algorithm-Voting EM in Personalized Education System, and We introduce Confidence Factor and Time Updating Mark to improve the efficiency and accuracy of the algorithm.Finally, we design and implement a Personalized Education System-Istudy, which uses LMFPE for learner modeling. Based on Istudy, this paper propose an experiment, which asks 52 students to learning ?Quadratic function? through Istudy and analyzes the distinguish between KMFPE and students? test score. Through the result of the experiment, we verify the validity of KMFPE and the accuracy of online evolution algorithm.