【摘要】 针对传统学生英语写作能力分类方法准确率偏低的情况,提出一种基于离散Hopfield神经网络(DHNN)的写作能力评价模型。首先,用层次分析法构建了学生英语写作能力评价指标体系,然后将写作能力分类指标划分为5个等级,进行编码,网络通过对分类标准的联想记忆,实现对于学生的英语写作能力的分类,并和BPNN模型的分类结果进行对比。仿真结果表明:BPNN模型分类准确率为80.0%,DHNN模型分类准确率为100.0%,DHNN模型提高了分类准确率和泛化能力且模型建立过程简单、结果直观,验证了所提出的模型的有效性。更多还原

【Abstract】 In view of the low accuracy of traditional methods in student English writing ability classification, an assessment model based on discrete Hopfield neural network(DHNN) is proposed. An assessment index system is first established using analytical hierarchy process. The indexes are divided into 5 classes and encoded. Then, the network classified student writing abilities through associative memory of classification standards, and the classification results are compared with those produced by a BPNN model. The simulation results show that first, the classification accuracy of the BPNN model is 80.0%, while that of the DHNN model is 100.0%; second, the DHNN model improves the classification accuracy and generalization ability; third, its establishment process is simpler and the output is more perceivable. Thus the effectiveness of the proposed model is verified.更多还原