数据驱动的精准化学习评价机制与方法  

Evaluation Mechanism and Method for Data-Driven Precision Learning

作　　者：黄涛[1] 赵媛 耿晶[4] 王涵 张浩 杨华利[2] HUANG Tao;ZHAO Yuan;GENG JING;WANG Han;ZHANG Hao;YANG Huali(不详)

机构地区：[1]华中师范大学人工智能教育学部、教育大数据应用技术国家工程实验室,湖北武汉430079 [2]华中师范大学国家数字化学习工程技术研究中心,湖北武汉430079 [3]九江学院,江西九江332005 [4]华中师范大学教育大数据应用技术国家工程实验室,湖北武汉430079

出　　处：《现代远程教育研究》2021年第1期3-12,共10页Modern Distance Education Research

基　　金：国家自然科学基金面上项目“面向时空融合的学习者认知诊断理论及关键技术研究”(619770332);中国科学院院士咨询项目“我国科学教育发展战略研究”(2018-Z10-A-025)。

摘　　要：综合素质评价是深化教育改革、落实立德树人根本任务的必然要求。信息技术赋能教育教学过程,使得传统教育评价机制向数据驱动的精准化学习评价方向迈进。数据驱动的精准化学习评价是一种不断获取、整合、分析学习过程中的多模态数据,对教学和学习活动的信息流与运行轨迹作出观测、解释和反馈,以发现教育中存在的问题,辅助课堂教学干预的评价方法,包含多场景数据采集、多空间数据融合、精准分析模型构建、分析结果可视化四个关键组成部分。多场景数据采集为学习评价提供数据来源和量化手段支撑,多空间数据融合为学习评价提供统一数据标准,精准分析模型为学习评价提供数据驱动的学习分析与评价方法,分析结果可视化为学习评价提供反馈和应用服务。数据驱动的精准化学习评价框架,应以区块链技术为基础,融合人工智能、云计算、学习分析、情境感知等新兴技术,在多维时空尺度上全面采集学习过程中的多元海量数据,通过全面、系统的统计分析和数据挖掘,从学习者的内在动机到认知发展、情感表现,再到社会综合交互能力,多维度、及时、准确地评估学生的学习状况。当前,探索数据驱动的精准化学习评价,可从构建符合多空间数据特征的综合素质评价框架,完善过程性数据的采集与共享机制,突破情感、认知、动机与社交评价的关键技术,以及研制数据驱动的学习评价平台等方面入手。Comprehensive quality evaluation is an inevitable requirement for implementing the fundamental task of deepening education reform and cultivating talents.Information technology promotes the process of education and teaching,which makes the traditional education evaluation mechanism move towards the direction of data-driven learning evaluation.Data-driven precision learning evaluation is an evaluation method that observes,explains and feedbacks on the information flow and running track of teaching and learning activities by continuously obtaining,integrating and analyzing multi-modal data in the learning process,so as to find problems existed in education and assist the intervention of classroom teaching.It contains multi-scene data collection,multi-spatial data fusion,accurate analysis model construction,and analysis result visualization.Multi-scene data collection provides data sources and quantitative means for learning evaluation;multi-spatial data fusion provides unified data standards for learning evaluation;accurate analysis model provides data-driven learning analysis and evaluation methods for learning evaluation and analysis results visualization provides feedback and application services for learning evaluation.The data-driven precision learning evaluation framework should integrate technologies such as artificial intelligence,cloud computing,learning analysis and context awareness based on the block chain technology.It can comprehensively collect multiple massive data in the learning process on multi-dimensional spatial-temporal scale.Through comprehensive and systematic statistical analysis and data mining,it can evaluate students’learning status from their intrinsic motivation,cognitive development,emotional performance,to comprehensive social interaction ability in a multi-dimensional,timely and accurate way.At present,to explore data-driven accurate learning evaluation,we can carry out research on building a comprehensive quality evaluation framework in line with the characteristics of multi-spat

关 键 词：数据驱动 精准化 学习评价 评价机制 评价方法 

分 类 号：G434[文化科学—教育技术学]