产教深度融合背景下《数据分析与挖掘》课程“AI赋能 + 真题真做”教学改革研究与实践

Research and Practice of “AI Empowerment + Real Problem Solving” Teaching Reform in the Course of Data Analysis and Mining under the Background of Deep Integration of Industry and Education

摘要: 随着生成式人工智能(Artificial Intelligence, AI)与大语言模型技术的发展，软件开发和数据分析正在形成以自然语言描述任务、由AI辅助完成代码生成与调试的人机协同编程方式。作为交通运输专业的重要交叉课程，《数据分析与挖掘》以语法记忆和代码复现为主的传统教学模式，难以充分满足应用研究型人才培养对业务理解、问题建模和工程实践能力的要求。本文围绕交通运输专业课程建设，提出“AI赋能 + 真题真做”教学改革方案：引入多源异构、含噪声和缺失值的真实交通数据，建设分层案例库；采用基于意图驱动编程的人机协同教学方法，将教学重点由单纯代码编写调整为交通业务分析、任务分解、提示词设计、代码审查和结果解释；以企业真实问题组织项目式实训，形成“企业出题、学生揭榜、校企双导师评价”的教学流程；建立过程性评价与项目评价相结合的考核体系。项目拟通过两轮教学实践，检验该方案对学生复杂工程问题求解能力、AI工具规范使用能力和工程表达能力的促进作用。

Abstract: With the development of generative artificial intelligence and large language models, software development and data analysis are increasingly adopting a human-AI collaborative workflow in which tasks are described in natural language and AI tools assist with code generation and debugging. As an important interdisciplinary course for transportation students, the traditional teaching model of Data Analysis and Mining, which emphasizes syntax memorization and code reproduction, cannot fully meet the requirements of applied research-oriented talent cultivation in business understanding, problem modelling, and engineering practice. This paper proposes an “AI empowerment + authentic problem solving” reform scheme. The scheme introduces real transportation data characterized by multi-source heterogeneity, noise, and missing values to construct a hierarchical case library; adopts an Intent-driven Programming-based human-AI collaborative approach that shifts the emphasis from code writing alone to transportation business analysis, task decomposition, prompt design, code review, and result interpretation; organizes project-based training around authentic enterprise problems through a process of enterprise problem setting, students taking up the challenges, and joint evaluation by university and enterprise mentors; and establishes an assessment system combining process evaluation with project evaluation. Two rounds of teaching practice are planned to examine the effects of the scheme on students’ ability to solve complex engineering problems, use AI tools appropriately, and communicate engineering results.