基于学生数字档案的自适应教学关键技术研究——面向项目式教学过程评价的学生、团队与岗位画像模型

Research on Key Technologies of Adaptive Teaching Based on Student Digital Profiles—Student, Team, and Role Profiling Models for Process Evaluation in Project-Based Learning

摘要: 高校项目式教学存在“重结果、轻过程”的评价困境，本文选取学院2025级的共165名学生作为研究对象。将学情统计、项目分组情况、作业时间、签到记录等多源数据进行整合，建立起包括112个字段的学生数字档案。从个人、团队和岗位三个层级提取学习投入、作业执行力、时间意识、课堂参与度、互动协作度、学习质量、团队健康度与岗位胜任力等指标，采用专家赋权与规则阈值相结合的方法构建学生画像、团队画像和岗位画像，并引入大语言模型基于脱敏后的结构化指标生成可解释性的教学诊断与干预建议。研究结果显示，该框架能够识别出成绩评价难以发现的协作薄弱、岗位虚化、团队差异过大等问题，为项目式教学的过程性评价和自适应教学提供数据支持。

Abstract: Project-based learning (PBL) in higher education is confronted with the evaluation dilemma of overemphasizing outcomes while neglecting processes. This paper takes a total of 165 students from the 2025 cohort of the college as research participants. By integrating multi-source data including learning status statistics, project grouping details, assignment completion duration and attendance records, a student digital profile system covering 112 fields is established. Indicators such as learning engagement, assignment execution capacity, time awareness, classroom participation, interactive collaboration performance, learning quality, team health and post competency are extracted from three levels: individual, team and post. A combined approach of expert weighting and rule-based thresholds is adopted to construct student portraits, team portraits and post portraits. Furthermore, a large language model (LLM) is introduced to generate interpretable teaching diagnosis and intervention suggestions based on desensitized structured indicators. The research findings demonstrate that the proposed framework can identify problems undetectable by traditional grade-based evaluation, such as weak collaborative competence, vague post responsibilities and excessive inter-team disparities. This study provides data support for the process-oriented evaluation and adaptive teaching of project-based learning.