分析挖掘教育大数据是智慧教育的必由之路。本项目针对“跨模态课堂大数据分析挖掘利用难题”，本研究旨在研究跨模态课堂数据中的教学活动要素的定义与形式化表述，识别教学活动要素，从微观和宏观层面深度挖掘与分析教学模式，建立描述教学过程和组织策略的教学模式库，提出一种科学、规范、有效、可计算的课堂教学质量评价方法，攻克教学活动要素动态性关联性、类型复杂多样、特征高维、值稀疏、评价来源多引起的教学活动与模式表示、多任务识别、相似度计算、多源评教融合决策的难题，并在西安交通大学本科教学中验证与示范。研究成果将有力支撑高校课堂教学质量科学评价与精准督导，促进形成大数据驱动的教学质量提升良性运行机制，驱动教育管理科学化。拟发表高质量国际期刊和会议10篇以上。

Analyzing and mining education big data is the inevitable way for achieving smart education. This project focus on the problem of ‘Mining, analysis and utilization of cross-modal classroom big data’. The study aims at studying the definition and formalization of teaching activity elements, identifying the elements of teaching activities, deeply mining and analyzing the teaching model from the micro- and macro- perspectives, establishing corresponding teaching model bases that describe the teaching process and organization strategies, and proposing a scientific, standardized, effective, and calculable classroom teaching quality evaluation method. It will overcome the difficulties of teaching activity and model representation, multi-task recognition, similarity calculation, and multi-source evaluation fusion decision caused by dynamic relevance of the elements of teaching activities, complexity and diversity of the types, high-dimensional features, sparse values and multisource-based evaluation and decision. It will be verified and demonstrated in the undergraduate teaching of Xi'an Jiaotong University. The research results will effectively support the scientific evaluation and precise supervision of classroom teaching quality in colleges and universities, and promote the formation of a benign operating mechanism of big data-driven teaching quality and drive the scientific management of education. This research proposes to publish above 10 pieces of high quality international journals and conferences papers.