摘要
评价学生的创造力是一项极富挑战性的工作,创造力评价是培养开发创造力的首要环节.当需要评价的被试、创造力指标个数、指标等级水平都较多时,创造力评价工作便陷入了僵局,利用人工甚至无法完成.学习矢量量化(LVQ)神经网络在模式识别方面具有良好的性能,据此尝试用LVQ神经网络进行创造力评价.根据创造力评价指标及等级的数目,构建了由输入层、隐含层、输出层组成的LVQ神经网络,用训练好的网络对测试样本进行仿真测试,仿真结果和实际情况正好相符,体现出LVQ神经网络在创造力评价中的实用和有效性.
Creativity assessment for students is a very challenging task, and it＇s the primary link in cultivating creativity. When the number of participants, creativity indices and index grade levels are greater, the task of creativity assessment will be an impasse, even uncompleted just artificial. LVQ has a good performance in pattern recognition. This paper attempts to use LVQ neural network for evaluation of creativity in view of the above. LVQ neural network consist of input layer, hidden layer and output layer is structured according to the number of creativity assessment index and its levels, and simulation testing is made for test samples with network trained well already. Simulation result and the actual situation are compatible perfectly, and this reflects the practicality and validity on creativity assessment for LVQ neural network.
出处
《数学的实践与认识》
CSCD
北大核心
2014年第16期288-292,共5页
Mathematics in Practice and Theory
基金
国家社会科学基金"十一五"规划课题(BBA080050)
关键词
创造力
智能评价
LVQ神经网络
训练
测试
creativity
intelligent evaluation
LVQ neural network
train
test
