联系我们

论文查询结果	返回搜索

论文编号：	11656
作者编号：	2120182908
上传时间：	2020/6/19 16:24:17
中文题目：	基于数据挖掘的在线旅游平台净推荐值预测研究
英文题目：	Research on NPS Value Prediction of OTA Platform based on Data Mining
指导老师：	安利平
中文关键字：	在线旅行；用户产生内容；净推荐值；数据挖掘；用户分类
英文关键字：	OTA;UGC;NPS;Data Mining;Customer Classification
中文摘要：	随着互联网的不断成熟，旅游行业也逐渐向线上化渗透，我国OTA（在线旅行）行业已有近20年的发展历史，新型的在线旅行产品愈发注重交互，加入了攻略、社区模块等，UGC（用户产生内容）模式应运而生，由于其内容更加全面和个性化，UGC模块成为OTA企业吸引用户的重要业务。而近年来经济下行，给旅游行业带来较大冲击，行业竞争也随之加剧，各OTA企业想要在竞争中占据有利地位，首先要吸引和保留更多忠实的用户，NPS（净推荐值）作为衡量用户忠诚度的指标受到更多企业的重视。各企业纷纷使用NPS工具进行用户调研，然而在互联网发展背景下，目前的NPS问卷主要通过网络平台进行线上发放，这其中存在诸多明显问题，导致企业获取到的用户忠诚度信息存在很大程度的失真。 针对这些问题，本研究获取了某OTA企业UGC模块的用户行为数据，结合过采样技术处理数据类别分布不平衡的情况，使用数据挖掘分类算法建立了NPS预测模型，对未填写问卷的用户进行NPS类型预测。实验结果证实了现有NPS在线调研存在的问题，对初始NPS值进行了修正，并挖掘预测出更多的贬损者和被动者，使企业可以更清晰地进行后续用户运营工作。 本研究的思路和方法还可以扩展应用于具有类似数据源的其他线上企业的调研过程，通过对线上NPS调研过程进行改进，可以进一步提高在线NPS调研的科学性及合理性，帮助管理者掌握真实的用户忠诚度情况，为企业提高用户忠诚度的工作提供数据指导。
英文摘要：	As the Internet matures, the tourism industry also gradually towards the online penetration, OTA (online travel) industry has a history of nearly 20 years, the new online travel products increasingly pay attention to interaction, thus the UGC (user generated content) model Arises. due to its comprehensive and personalized content, UGC module has become an important business for the OTA enterprises to attract users. In recent years, the economic downturn has brought a great impact to the tourism industry, and the industry competition has also intensified. In order to occupy a favorable position in the competition, OTA companies should first attract and retain more loyal users. As an index to measure user loyalty, NPS (net recommendation value) has attracted more attention from enterprises. Various enterprises have been using NPS tools to conduct user surveys. However, the current NPS questionnaire is mainly distributed online, which has many obvious problems, resulting in a large degree of distortion of user loyalty information obtained by enterprises. In view of these problems, this study acquired user behavior data of an OTA enterprise UGC module, combined with oversampling technology to deal with the unbalanced distribution of data categories, and established an NPS prediction model by using data mining classification algorithm to predict NPS types of users who did not fill in the questionnaire. The experimental results confirmed the problems existing in the existing NPS online survey, corrected the initial NPS value, and mined more derogators and passives, so that the enterprise can carry out the follow-up user operation work more clearly. The ideas and methods of this study can also be extended to other online businesses with similar data sources of research process, through the online NPS investigation process was improved, can further improve the scientificity and rationality of NPS survey online, to help managers to grasp the real user loyalty, for enterprises to improve customer loyalty to provide data guidance work.
查看全文：	预览  下载(下载需要进行登录)