摘要
在线评论聚集了海量意见与建议,可以为物流企业的运营管理提供方向性参考。文章提出基于情感分析的用户创意挖掘方法,帮助物流运营管理获得具有创新性的用户反馈,以提升运营效率。文章以邮政快递为研究对象,以抖音和快手短视频物流评论为数据源,通过爬虫获取数据;其次利用改进的LSTM情感分析模型Bi-LSTM实现情感分类;最后以词云聚焦话题点,以语义网络分析图并结合Apriori关联规则分析算法可视化出话题背后最后一公里的服务问题,以便资源调整与服务经营的改进。实验结果表明,用户对于快递员很少送货上门转而放入快递柜不满;在物流末端的偏远地区配送服务一般集中在乡镇地区,农村配送难以实现;用户对于贵重物品损坏丢失以及快递员服务态度反映强烈。
User online reviews gather a huge amount of opinions and suggestions,which can provide directional references for the operation and management of logistics enterprises.This paper proposes a user creativity mining method based on sentiment analysis to help logistics operation management obtain innovative user feedback and improve operational efficiency.This study takes postal logistics as the research object,takes Tiktok and Kuaishou video comments as the data source,and obtains data through crawlers;then the improved LSTM sentiment analysis model Bi-LSTM is used to achieve sentiment classification;finally,word cloud is used to focus on topic points,and the network semantic graph is combined with Apriori association rule analysis algorithm to visualize the service problem behind the last mile of the topic,in order to adjust resources and improve service management.The results of the experiment show that users are dissatisfied with the fact that couriers rarely deliver to their homes and put them in the delivery lockers instead.Distribution services in remote areas at the end of logistics are generally concentrated in township areas,and rural distribution is difficult to achieve.Users strongly reflect on the loss of damaged valuables and couriers service attitude.
作者
李浩洋
张瑞军
周瑜
LI Haoyang;ZHANG Ruijun;ZHOU Yu(School of Management,Wuhan University of Science and Technology,Wuhan 430065,China;Center of Service Science and Engineering,Wuhan University of Science and Technology,Wuhan 433065,China)
出处
《物流科技》
2024年第11期64-67,共4页
Logistics Sci-Tech
基金
武汉科技大学研究生创新创业基金项目(JCX2021039)
湖北省高等学校省级教学研究项目(2020369)
湖北省高等学校哲学社会科学研究重点项目(21D014)。
关键词
物流评论
服务改进
创意挖掘
情感分类
关联分析
logistics reviews
service improvement
creativity mining
sentiment classification
association analysis