PCA和Elman网络在移动学习策略分类中的应用  

Application of PCA and Elman network in mobile learning strategy classification

作　　者：胡帅[1] 程迎新[1] 顾艳[1] 

机构地区：[1]渤海大学大学外语教研部,锦州121013

出　　处：《电子测量技术》2016年第5期182-186,共5页Electronic Measurement Technology

基　　金：辽宁省教育厅科学研究一般项目(W2015015);辽宁省社会科学基金(L14CYY022);辽宁省社会科学基金重点项目(L15AYY001)资助

摘　　要：针对传统的大学生英语移动学习策略分类方法准确率较低的情况,提出了一种主成分分析(PCA)和Elman神经网络相结合的分类模型。首先,用PCA对所获得的移动学习策略原始数据作数据降维处理,提取前5个主成分,建立新的特征样本矩阵,再对Elman神经网络进行训练和泛化能力测试。仿真结果表明:单一的BPNN分类准确率为70.0%,单一的Elman网络分类准确率为80.0%,PCA-Elman网络分类准确率为100.0%,PCA-Elman网络模型简化了单一Elman网络的结构,提高了网络的训练速率、分类准确率和泛化能力,验证了所提出的模型的有效性。To overcome the problem of low accuracy of traditional methods in the area of college student mobile learning strategy classification,a new classification model based on principal component analysis（PCA）and Elman neural network is proposed.First,dimensionality reduction was done to the obtained original data of student mobile learning strategies using PCA and 5principal components were extracted to create a new feature sample matrix.Then the Elman neural network was trained and its generalization performance was tested.The simulation results indicate that：the classification accuracy of the single BPNN is 70.0%,the one of the single Elman model is 80.0% and the one of the PCAElman model is 100.0%;the PCA-Elman model can simplify the structure of the single Elman network,improve the training speed,classification accuracy and generalization performance;the effectiveness of the recommended model is proved.

关 键 词：主成分分析 ELMAN神经网络 BP神经网络 移动学习策略 分类 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]