摘要
酒店顾客满意度作为一项重要指标,是顾客决策与酒店管理行为的重要依据。本文对酒店预订平台的在线评论进行情感分析获取顾客满意度,分别采用Word2vec,GloVe,fastText,BERT预训练词向量作为模型词嵌入层,并与CNN,LSTM等模型进行对比分析得出最优模型。本文选取携程网站上福州市内多家知名酒店的在线评论实例论证,实验结果显示:BERT-BiLSTM模型准确率达85.8%。本文探究了各酒店的顾客满意度水平,为顾客选择酒店的决策行为以及福州市内知名酒店的发展提供参考依据。
As an important indicator,hotel customer satisfaction is an important basis for customer decision-making and hotel management behavior.This paper conducts emotional analysis on online reviews of hotel reservation platform to obtain customer satisfaction.Word2vec,GloVe,fastText,BERT pre-training word vectors are respectively used as the model word embedding layer,and compared with CNN,LSTM and other models to obtain the optimal model.This paper selects online reviews of many well-known hotels in Fuzhou on Ctrip’s website to demonstrate.The experimental results show that the BERT-Bi LSTM model has an accuracy rate of 85.8%.This paper explores the level of customer satisfaction in each hotel,providing a reference for customers’decision-making behavior in choosing hotels and the development of well-known hotels in Fuzhou.
作者
高丽君
张宇涛
林昀萱
施慧玲
GAO Li-jun;ZHANG Yu-tao;LIN Yun-xuan;SHI Hui-ling(College of Economics and Management,Fuzhou University,Fuzhou 350108 China;College of Mathematics and Statistics,Fuzhou University,Fuzhou 350108 China;Maynooth International Engineering College,Fuzhou University,Fuzhou 350108 China)
出处
《科技创新与生产力》
2022年第12期65-70,共6页
Sci-tech Innovation and Productivity
基金
福州大学国家级大学生创新创业训练计划项目(202110386018)。
关键词
在线评论
情感分析
酒店顾客满意度
词向量
双向长短期记忆网络
online reviews
emotional analysis
hotel customer satisfaction
word vector
bidirectional long and short term memory network(BiLSTM)
