题名 |
从搜索序列文本看高端商务车 |
并列篇名 |
Sequence Text from the search to see the high-end commercial vehicl |
DOI |
10.6338/JDA.201606_11(3).0002 |
作者 |
朱雪宁(Hsueh-Ning Chu);王汉生(Han-Sheng Wang) |
关键词 |
搜索引擎数据 ; 文本分析 ; Logistic回归模型 ; 品牌忠诚度 |
期刊名称 |
Journal of Data Analysis |
卷期/出版年月 |
11卷3期(2016 / 06 / 01) |
页次 |
17 - 36 |
内容语文 |
簡體中文 |
中文摘要 |
本文对100 万搜索引擎用户的13 亿搜索序列文本进行探索分析,对高端车用户以及商学院人群做了描述对比,并针对用户搜索高端车品牌过程中的动态选择行为进行建模。首先,我们发现,在人群划分上,高端车用户和商学院用户表现出更加高端的属性,这主要表现在他们对生活、事业、学业上更卓越的追求。接下来,本文利用逻辑回归构建了忠诚模型,对用户在搜索过程中表现的忠诚和叛变行为进行了刻画,并对影响其忠诚行为的关键因素进行了逐一分析。根据模型的估计结果,我们发现,用户的搜索时间间隔、搜索关键词长度、搜索点击数等指标对用户忠诚行为有显著影响3。最后,我们利用成本收益曲线对模型进行了评价,并得到了良好的效果。 |
英文摘要 |
In this paper, 1.3 billion text search sequence 1 million search engine users to explore the analysis of high-end car users and business school population was described comparison, and search for the high-end car brand in the process of modeling the dynamic behavior of choice for users. First, we found that the population is divided, high-end car users and business schools exhibited more high-end user property, which is mainly manifested in their previous life, career, superior academic pursuit. The next key factor, we use logistic regression models to build loyalty, loyalty and treachery users in the search process performance were portrayed, and affect their loyalty behavior were analyzed one by one. According to the estimated results of the model, we found that users of search time interval, the length of a keyword search, the search hits and other indicators have a significant impact on customer loyalty behavior. Finally, we use the cost of revenue curve model were evaluated and obtained good results. |
主题分类 |
基礎與應用科學 >
資訊科學 基礎與應用科學 > 統計 社會科學 > 管理學 |
参考文献 |
|