摘要： 针对协作学习中基于学习者特征的分组方式对学习过程的影响,设计一种基于改进细菌觅食的协作学习分组算法。在实现协作学习分组过程中,引入分组调节因子和特征权值,满足不同教学活动对学习者多个特征及分组的要求。为构成有效的分组空间,在细菌种群初始化中,细菌群体以实数编码,并加入随机扰动以增加细菌种群的多样性;在算法后期加入二次变异操作,以避免细菌觅食算法可能出现的早熟收敛现象。仿真实验结果表明,该算法在不同分组形式下,与传统算法相比,具有较优的分组性能和较高的准确率,并且对于不同数据集规模具有良好的稳定性。
关键词:
协作学习,
评价准则,
学习分组,
分组形式,
多目标优化,
细菌觅食优化算法
Abstract: The grouping form for collaborative learning group based on learners’ characteristics is one of the factors
that enhance the learning effectiveness． A new learning grouping algorithm based on enhanced bacterial foraging is proposed． In order to meet the requirements of different learning activity that is associated to the learners’ characteristics, the regulatory factor and feature weights are used to grouping． At the initialization step of algorithm,there are two method which are used to ensure the effective grouping space,one is that the bacterial population is coded by real number coding, and another is that a random perturbation is used to increase the diversity of bacterial populations． And the algorithm is
joined the second mutation to avoid the premature convergence at the later step． Simulation experimental results show that the proposed algorithm is advantage to increase the effectiveness and the accuracy for grouping form． And it has a good stability for data sets with different sizes
Key words:
collaborative learning,
evaluation criteria,
learning grouping,
grouping form,
multi-objective optimization,
bacterial foraging optimization algorithm
中图分类号: