我国在2001年《基础教育课程改革纲要（试行）》和2010年《国家中长期教育改革和发展规划纲要（2010-2020）》中强调要将个性化教学作为教育改革与发展的重要内容，以顺应新时代对创新型人才的诉求。学习者建模是个性化教学的基础和核心，通过挖掘学习者的相关数据，分析学习者的特征，能够帮助教师深入了解学习者，从而更好地满足学习者的个性化需求。然而已有研究指出现有的学习者建模过程中存在了很多问题，如缺少对学习者模型基础元素—学习者信息的细化分析，忽略了学习者特征维度之间的关系，因此构建结构清晰、要素完备的学习者建模过程框架成为当前需要重点解决的问题。 本研究基于学习者模型及学习分析的相关研究，总结和归纳出学习者建模的过程框架，为学习者模型的构建提供了方法层面的借鉴，该框架共有四个层次：逻辑层、数据层、分析层和应用层。逻辑层主要是围绕应用层的目的与需求，基于具体的学习情境及在线学习的理论，确定需要考察的理论模型。数据层主要发挥收集数据、处理数据的功能，为分析层提供可以直接输入的特征数据。分析层为核心模块，在教育学和数据科学理论的指导下为逻辑层和数据层建立双向联系，从数据层选取与教学意图、理论模型相适切的特征数据，并结合相应的算法来构建全面、具有可行性的数学模型。应用层提出了建模的需求，是学习者模型的根本，决定了模型的价值和意义所在，为模型的构建指明了方向，该建模过程框架虽然呈现出线性效果，但在实际应用中是循环迭代的过程，通过应用层对模型的检验后将结果反馈到逻辑层与数据层，从而进一步调整和优化。 为了验证该建模框架的有效性，本研究以奥鹏教师网络研修平台为例开展实证研究，首先通过理论演绎和专家访谈法构建了网络研修环境下的学习者理论模型，该理论模型包括预设资源利用度、生成性资源利用度、协作交流度、任务参与度和成果质量五个维度。其次基于平台数据结合学习者规范标准全面设计了该情境下的数据汇聚库与特征体系，然后通过专家和主成分分析两种方法的结合构建了数学模型，基于该学习者模型实现了学习者分类和个性化推荐系统，能够满足教师和管理者的多样化需求，有助于奥鹏教师培训平台提高培训的针对性和实效性，使教师培训更加适应新的需求和形势。

In 2001, China's "basic education curriculum reform outline (Trial)" and "the national medium and long term education reform and development plan (2010-2020)" in 2010 emphasize the importance of individualized teaching as an important content of educational reform and development, in order to comply with the demands of the new era for innovative talents. In the course of the development of individualized teaching concept, the emergence and popularization of information technology provides a powerful material basis and technical support, and has promoted a more accurate and personalized teaching. Learner modeling is the foundation and core of individualized teaching. By digging out the relevant data of the learners and analyzing the characteristics of the learners, they can help the teachers to understand the learners in depth and better meet the individual needs of the learners. However, there have been many problems in the process of learner modeling, such as the lack of a thinning analysis of the learner's model basic element - learner information. The model is not based on any learning and teaching theory and neglects the relationship between the learners' characteristic dimensions, so the construction is clear and the elements are complete. The modeling process framework has become a major problem to be solved. Based on the study of learner model and learning analytics, this research summarizes and summarizes the process framework of learner modeling, which provides a reference for the construction of learner model. The framework consists of four levels: logic layer, data layer, analysis layer and application layer. The logic layer mainly focuses on the purpose and requirements of the application layer, and determines the theoretical model to be investigated based on the specific learning context and online learning theory. The data layer mainly functions to collect data and process data, and provides feature data that can be directly input to the analysis layer. The analysis layer is the core module. Under the guidance of the theory of pedagogy and data science, it establishes a two-way connection for the logic layer and the data layer, selects the characteristic data which is appropriate to the teaching intention and the theoretical model from the data layer, and constructs a comprehensive and feasible digital model with the corresponding algorithm. The application layer puts forward the requirement of modeling, which is the root of the learner model, determines the value and significance of the model, and indicates the direction for the model construction. Although the model presents a linear effect, it is a cyclic iterative process in the practical application, and the result is fed back to the logical layer and number after the application layer is tested for the model. According to the layer, it will be further adjusted and optimized. In order to verify the validity of the modeling framework, this study takes openg teachers' network research platform as an example to carry out an empirical study. First, the theoretical model of the learner is constructed through theoretical deduction and expert interview. The theoretical model includes the utilization degree of preset resources, the utilization of generative resources, and the degree of cooperation and communication. Five dimensions of the degree of task participation and the quality of the results. Secondly, based on the platform data and the standard standard of the learner, the data aggregation library and the feature system are designed in this context, and then the mathematical model is constructed by combining the two methods of expert and principal component analysis. Based on this model, the learners' classification and personalized recommendation system can be realized, which can satisfy the teachers and managers. The demand for sampling helps to improve the pertinence and effectiveness of the training platform, so that teacher training can better adapt to the new needs and situations.