智能学习系统中学习习惯建模的方法研究

Research on Method of Modeling Learning Habits in Intelligent Learning System

学习习惯建模是精准描绘学习者肖像的重要部分,是设计智能学习系统亟须解决的问题。然而,由于学习习惯的形成具有多源性,其应用具有多层面性,导致学习习惯测量模型具有多场景、多学段的特征,以及面向实践的理论选择的特征。基于这些特征,单一学科的方法都无法系统地解决学习习惯建模的问题,研究通过探索教育的常识性概念与学习行为数据建立关联的可行路径,采用混合式研究思路设计了一套可行的、可复用的学习习惯综合建模方法,该方法可以让研究者根据自己的研究场景来构建学习习惯测量模型,从而实现计算机的自动诊断。学习习惯综合建模方法将教育学领域的观察法和问卷调查法、心理学领域的反应频率法以及计算机领域的数据挖掘法相结合,包括测量模型维度的构建、测量模型指标的选择、模型特征值的计算、模型的生成四个过程。

Constructing a model of learning habits is an important part of accurately portraying the learners,and an urgent problem to be solved in the design of intelligent learning system.However,since the formation of learning habits is multi-source and its application is multi-faceted,the measurement model of learning habits has the characteristics of multi-scene,multi-segment and practice-oriented theoretical selection.Based on these characteristics,a single disciplinary approach cannot solve the problem of modeling learning habits systematically.Through exploring the feasible ways to establish the association between commonsense concepts of education and the learning behavior data,this study adopts the hybrid research approach to design a set of feasible and reusable comprehensive methods to modeling learning habits.This method supports researchers to construct the model of learning habit measurement according to their own research scenarios,so as to realize the automatic diagnosis by computer.This method combines the observation method and questionnaire survey in the field of education,the response frequency method in the field of psychology with the data mining method in the field of computer.And it includes four processes of the construction of measurement model dimension,the selection of measurement model index,the calculation of model eigenvalue and the generation of model.