基于GloVe-BiLSTM的在线研讨信息分类模型研究  

Research on Online Discussion Information Classification Model Based on GloVe-BiLSTM

作　　者：易明[1] 李藿然 刘继月 Yi Ming

机构地区：[1]华中师范大学信息管理学院,湖北武汉430079

出　　处：《情报理论与实践》2022年第9期173-179,172,共8页Information Studies:Theory & Application

基　　金：国家社会科学基金重点项目“在线健康社区知识共创机理及引导机制研究”的成果之一,项目编号:21ATQ006。

摘　　要：[目的/意义]把握并识别在线研讨信息类型,有助于研讨用户快速获取研讨信息、减轻认知负担、形成对研讨状态的正确判断,从而提升研讨效率。[方法/过程]基于深度学习的文本语义理解和挖掘,构建在线研讨信息分类识别的深度学习组合模型GloVe-BiLSTM,进行在线研讨信息的自动化分类预测。利用GloVe对待分类文本进行训练,以获得字词级别的向量,再将词向量输入BiLSTM层提取语义特征,最后输入新的在线研讨文本以得到分类预测的最终结果。[结果/结论]针对CMV社区在线研讨信息的实验表明,构建的GloVe-BiLSTM组合模型在分类准确率、精确率、召回率、F1值等方面均具有出色表现,能够有效实现在线研讨信息的类型分类,为在线研讨社区服务优化提供参考。[Purpose/significance]It is helpful for users to identify and judge the type of online discussion,so as to reduce the burden of online discussion and help users to obtain information quickly.[Method/process]Based on the text semantic understanding and mining of deep learning,this paper constructs the deep learning combination model GloVe-BiLSTM for the classification and recognition of online discussion information,and carries out the automatic classification and prediction of online discussion information.GloVe is used to train the classified text to obtain the word level vector,then input the word vector into the BiLSTM layer to extract semantic features,and finally input a new online discussion text to obtain the final result of classification prediction.[Result/conclusion]The experiment of online discussion information in CMV community shows that the GloVe-BiLSTM combination model has excellent performance in classification accuracy,precision,recall and F1 value.It can effectively realize the type classification of online discussion information and provide reference for the optimization of online discussion community service.

关 键 词：在线研讨信息 深度学习 BiLSTM模型 GloVe模型 信息分类 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]