项目式学习是一种能够促进学生自主学习，并激发学生深度学习的学习模式。随着教学改革的实行，其越来越受到各个中小学的重视。然而已有的研究表明，学生很难适应这种以学生为中心的学习方式，常常出现学习目标偏离，学习策略模糊等问题，而教师又无法时时注意每个学生的具体学习情况，并给予足够指导。. 为了解决项目式学习实行中的问题，本项目提出利用教学辅助系统相关技术，主动采集学生项目式学习过程；以学习行为挖掘、自然语言处理相关技术，分析辨别学生高效和低效的个人以及协作学习策略，实现对学习过程的重要特征提取；并在此之上，利用贝叶斯学生能力模型，有限状态机的相关技术，构建项目式学习的学生模型，追踪学生的学习状态；并最终依据学生模型，实现辅导干预，以提高项目式学习的效果。

Project-based learning (PBL) can urge students to self-learn and leverage deep learning. With the development of the educational policy issued by the Ministry of Education, more and more elementary schools put high value on PBL. However, the previous studies showed that students had difficulties in adapting to this student-centered learning style. They often lost the learning objects and took inefficient learning strategies. On the other hand, the teachers often cannot give every student enough facilitation, given the size of a normal class. ..To deal with the issues mentioned above, this research proposal try to use the technique of learning management system to record students' learning procedure data; to use educational data mining and natural language processing to classify efficient and inefficient learning strategies; then, construct the corresponding student models with the help of Bayesian knowledge tracing and finite state machine to track students' learning states; design and implement appropriate learning interventions in the final.