形成性测评系统自动组卷策略实现

Realization of automatic test paper strategy in formative evaluation system

形成性测评是指对学习者学习过程的全面测评,对学习者课程学习成果的阶段性考核。在线考试是远程教育形成性测评中最基本、最常用的方式之一,编制高质量的试卷可以有效地检验学生对知识的掌握水平及教师的教学效果。遗传算法是建立在自然选择和群体遗传学基础上的一种计算优化方法,适合于并行处理,具有应用范围广、收敛速度快等显著特点。而形成性测评中自动组卷策略是需要教师根据不同教学阶段、不同教学任务、不同学生学习状况进行多方面的组卷要求的过程,这是一个多目标优化求解问题的过程。采用遗传算法进行自动组卷设计,从参数编码、适应函数的设计、遗传操作的设计、控制参数的设计等几个方面进行设计和改进,系统可自动生成多份按策略规划的试卷,保证了试卷的准确率和科学性。

Formative evaluation refers to a comprehensive evaluation of learners learning process;it is the periodic assessment of learners＇ curriculum learning outcomes.The online exam is one of the most basic,most common ways in formative evaluation of distance education,and the composition of high-quality examination papers can effectively test the effect of the students＇ mastery level of knowledge and teachers＇ teaching effectiveness.The genetic algorithm is an optimum calculating method on the base of natural selection and population genetics,is suitable for parallel processing with a wide range of applications,high convergence speed and other notable features.While the automatic test paper strategy in formative evaluation is a process that needs teachers meeting various requirements of generating test paper according to the different stages of teaching,different teaching tasks and different students＇ learning situations,consequently,this is a process of solving multi-objective optimization problems.Using genetic algorithms for automatic generating test paper design,it designs and makes improvement from the aspects of encoding parameter,adapting to the design of the function,the genetic manipulation,the control of parameter etc.,and the system can automatically generate multiple copies of papers that based on the strategic planning,so as to ensure the papers＇ accuracy and scientific.