摘要
针对当前线上学习平台在教学质量评估分析的过程中存在分类效果差、预测准确率较低的问题,提出了一种利用深度学习框架对多源数据进行分析与处理的算法。其使用带文本联系算法的TextRank进行多源数据处理,以获得带有权值的词向量。并将词向量输入至基于CNN-BiGRU的学习行为特征提取模型中,利用CNN和BiGRU的局部及全局特征提取功能进行分析。同时将训练完毕的特征向量输入至Softmax模型进行评分,从而得到学习质量的评估值。在实验测试中,所提算法的精确率可达78.54%,且综合性能指标较为理想,证明其可有效提升线上平台学习质量评估的准确性。
In view of the problems of poor classification effect and low prediction accuracy in the evaluation and analysis of teaching quality on the current online learning platform,an algorithm to analyze and process multi-source data using deep learning framework is propoesd.The algorithm uses TextRank with text association algorithm for multi-source data processing to obtain the word vector with weight.Input the word vector into the learning behavior feature extraction model based on CNN-BiGRU,analyze it by using the local and global feature extraction functions of CNN and BiGRU,and input the trained feature vector into Softmax model for scoring,and get the evaluation value of learning quality.In the experimental test,the accuracy rate of the proposed algorithm can reach 78.54%,and the comprehensive performance index is ideal,which proves that it can effectively improve the accuracy of online platform learning quality evaluation.
作者
朱琪
ZHU Qi(Xi'an Aeronautical Polytechnic Institute,Xi'an 710089,China)
出处
《电子设计工程》
2023年第23期85-89,共5页
Electronic Design Engineering
基金
陕西省教育科学“十四五”规划2021年度课题(SGH21Y0564)。
