Grouping Students for Maximizing Learning from Peers
Grouping Students for Maximizing Learning from Peers
复制标题
对学生进行分组，以最大限度地向同龄人学习
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
Narasimha Murty Musti
中科院分区：
文献类型：
--
作者：
R. Agrawal;Sharad Nandanwar;Narasimha Murty Musti
我们研究将一个班级的N名学生分成k组，每组N名学生（N = k × N），使他们从同伴互动中学习的能力最大化的问题。在我们对这一问题的形式化表述中，任何学生都能够将自己所学课程的分数提高到与能力较高的同龄人中处于p百分位的学生的分数相同的水平。相比之下，过去的工作假设只有分数低于小组平均水平的学生才能提高他们的分数。我们给出了一个分区算法，使所有学生在任意p值下的总增益最大化，使得100 /(100−p)为整数值。该算法的时间复杂度仅为O （N log N）。我们还提出了使用实际数据的实验结果，表明所提出的算法优于当前策略。
We study the problem of partitioning a class of N students into k groups of n students each ( N = k × n ), such that their learning from peer interactions is maximized. In our formalization of the problem, any student is able to increase his score in the subject the class is studying up to the score of the student who is at p -percentile among his higher ability peers. In contrast, the past work presumed that only students with score below the group mean may increase their score. We give a partitioning algorithm that maximizes total gain summed over all the students for any value of p such that 100 / (100 − p ) is integer valued. The time complexity of the proposed algorithm is only O ( N log N ). We also present experimental results using real-life data that show the superiority of the proposed algorithm over current strategies.