“互联网+”环境下融合多特征的弹幕数据情绪监测模型构建  

Construction of Multi-Featured Barrage Data Sentiment Monitoring Model Under the"Internet+"Environment

作　　者：刘家威 郑军红[1] 何利力[1] LIU Jiawei;ZHENG Junhong;HE Lili(School of Computer Science and Technology(School of Artificial Intelligence),Zhejiang Sci-Tech University,Hangzhou 310018,China)

机构地区：[1]浙江理工大学计算机科学与技术学院(人工智能学院),浙江杭州310018

出　　处：《软件导刊》2024年第2期17-24,共8页Software Guide

基　　金：浙江省重点研发“尖兵”攻关计划项目(2023C01119)。

摘　　要：针对企业营销活动中未充分考虑客户情感特征,导致营销效果达不到预期的问题,提出一种融合多特征的ALBERT-SA-BIGRU模型。首先,基于企业营销活动弹幕数据构建表情符号词典及相关语料库。其次,将弹幕文本与弹幕属性共同输入ALBERT模型中提取弹幕文本的特征表示,并与GloVe预先训练好的表情符号特征融合。再次,利用自注意机制捕捉表情符号与弹幕文本、弹幕属性之间的关系,将捕获的词语特征输入BiGRU中从前向、后向捕获信息,加强语义依赖,提取情感特征。最后,用Softmax逻辑回归进行情感倾向性分类,构建情绪监测图。以某互联网营销平台163253条弹幕数据对模型进行性能验证的结果表明,模型准确率、精确率、召回率分别为88.8%、88.7%、88.9%,相较于其他模型均有一定提升,可为企业在营销活动中实现用户情绪监测智能化精准营销提供支持。A multi feature fusion ALBERT-SA-BIGRU model is proposed to address the issue of insufficient consideration of customer emo⁃tional characteristics in enterprise marketing activities,resulting in unsatisfactory marketing outcomes.Firstly,based on the barrage data of enterprise marketing activities,construct an emoticon dictionary and related corpus.Then,the bullet text and bullet attributes are jointly input into the ALBERT model to extract the feature representation of the bullet text,and fused with the pre trained emoji features of GloVe.Next,us⁃ing self attention mechanism to capture the relationship between emoticons,bullet text,and bullet attributes,the captured word features are input into BiGRU to capture information in both forward and backward directions,strengthen semantic dependencies,and extract emotional features.Finally,use Softmax logistic regression to classify emotional tendencies and construct an emotional monitoring graph.The perfor⁃mance verification of the model using 163253 bullet screen data from a certain internet marketing platform shows that the accuracy,preci⁃sion,and recall rates of the model are 88.8%,88.7%,and 88.9%,respectively.Compared with other models,the model has improved to a certain extent and can provide support for intelligent and precise marketing of user sentiment monitoring in marketing activities for enterprises.

关 键 词：多特征 ALBERT模型 GloVe模型 自注意机制 BIGRU模型 情绪监测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]