航空公司微博评论的意见信息抽取研究——以国航、南航和东航为例

发布时间：2018-04-24 17:57

本文选题：意见信息 + 条件随机场　；参考：《系统科学与数学》2017年04期

【摘要】：如今越来越多的乘客选择乘坐舒适快捷的飞机出行,中国航空运输需求因此逐年增长,航空公司在获得更多盈利空间的同时也面临激烈的竞争.对航空公司的用户评论进行意见信息抽取,不仅可用于航空公司改进服务质量和用户体验,还可为用户选择满意的航空公司提供参考.文章首次以新浪微博平台上航空公司的用户评论为基础数据,利用条件随机场进行意见信息抽取.在有关研究中,专家学者大多凭借以往知识的了解对特征对象和特征词进行人工标注,鲜少分析用户在本评论语料中的关注点.因此,文章创新性地在人工标注前首先利用TF-IDF算法进行关键词提取,找到本评论语料中用户的关注点,最后以超过93%的F平均值证明模型的有效性,为后续的研究提供了新方向.
[Abstract]:Nowadays, more and more passengers choose to travel by comfortable and fast air. As a result, the demand for air transportation in China is increasing year by year, and airlines are facing fierce competition while gaining more profit space. It can not only be used to improve the service quality and user experience of airlines, but also provide reference for users to select satisfactory airlines. Based on the comments of airline users on Sina Weibo platform for the first time, the paper uses conditional random field to extract opinion information. In the relevant studies, most experts and scholars rely on the knowledge of the past to manually annotate feature objects and feature words, and rarely analyze the concerns of users in this review corpus. Therefore, this paper innovatively uses TF-IDF algorithm to extract keywords before manual annotation, finds the user's concerns in this review corpus, and finally proves the validity of the model with more than 93% F average. It provides a new direction for further research.
【作者单位】：东北财经大学统计学院;辽宁大学亚澳商学院;
【基金】：国家社科基金重大项目(13&ZD171) 辽宁省教育厅科学研究项目(LN2016YB026) 东北财经大学研究生教学改革研究项目(yjyb201634)资助课题
【分类号】：TP391.1

【相似文献】