牟智佳， 王卫斌， 李雨婷， 严大虎
【关 键 词 】：
MOOCs； 个性化学习需求； 预测建模； 系统动力学； 仿真分析
The prediction of personalized learning needs based on learner's early learning performance can further optimize learner's learning experience and improve his participation in curriculum under the MOOCs environment. This study takes the learner's personalized learning needs as the research content, adopts the system dynamics as the guiding method, and cross uses the analytic hierarchy process and nonlinear regression analysis to determine the quantitative relationship between variables to establish a prediction model for personalized learning needs. Finally, the simulation analysis is carried out with the data of two courses in different languages, the changes of learners' learning needs in various aspects and the highly leveraged factors causing the changes of learning needs are explored and verified. The research results show that the prediction model consists of four subsystems including content, resources, process and evaluation. The model covers three state variables, four flow variables, twenty-three auxiliary variables and twenty constants, which can accurately predict the personalized learning needs of learners. The demand of content difficulty and evaluation standard are the main embodiment of personalized learning needs, which are positively correlated with the total amount of knowledge and the total amount of learning input of learners. Learning interest, demand satisfaction, and curriculum objectives are the high leverage factors that need to be paid attention to in prediction. In different courses, the main embodiment of the changes of learners' personalized learning needs and the high leverage factors that need to be paid attention to are the same. However, there will be a change with the high leverage factors' influence degree in different courses.