| 基于行为分析的Web日志用户识别算法 |
| |
| 引用本文： | 汤伟,黄培磊,陈璐艺,林祥.基于行为分析的Web日志用户识别算法[J].软件产业与工程,2013(6):53-56. |
| |
| 作者姓名： | 汤伟 黄培磊 陈璐艺 林祥 |
| |
| 作者单位： | [1]盘石软件上海有限公司,上海200333 [2]上海交通大学电子信息与电气工程学院,上海200240 [3]上海交通大学信息安全工程学院,上海200240 |
| |
| 基金项目： | 上海市科委科研计划项目（编号11DZl1501900）. |
| |
| 摘 要： | 文章提出了一种Web日志分析中数据预处理阶段用户识别的新算法。这种算法基于用户浏览行为的建模，能够很好地处理同一用户在不同地点上网情况下的用户识别问题。本算法通过归一化的支持度指标来选择最能代表用户浏览行为的行为模式，并用布尔编码的方式提取特征。用户间的相似度采用余弦相似方法计算，并利用KNN（K=I）的分类方法来识别用户。在特定构造的数据集上的实验表明文章提出的算法能够不依赖于IP信息识别用户并具有较好的识别率。
|
| 关 键 词： | 行为分析Web日志分析用户识别 |
Web Log User Identification Based on Behavior Analysis |
| |
| TANG Wei,HUANG Peilei,CHEN LuyF LIN Xiang.Web Log User Identification Based on Behavior Analysis[J].Software Industry and Engineering,2013(6):53-56. |
| |
| Authors: | TANG Wei HUANG Peilei CHEN LuyF LIN Xiang |
| |
| Affiliation: | 3 l(Pansafe Software, Shanghai 200333, China) 2(Department of Electrical Engineering, Shanghai Jiaotong University, Shanghai 200240, China) 3(School of Information Security Engineering, Shanghai Jiaotong University, Shanghai 200240, China) |
| |
| Abstract: | This paper proposed a new algorithm for user identification in preprocessing stage of web log analysis. he algorithm is based on model of user browsing behavior, and performs well in dealing with situation of user accessing the website from different locations. Features are selected with normalized support and are binary coded. Similarity between users is calculated with cosine similarity measure, and K-nearest neighbor （K=I） is used as classification algorithm. In a sDeciallv constructed dataset, the proposed algorithm performs well in identifying users without relying on IP addresses. |
| |
| Keywords: | Behavior Analysis Web log Analysis User Identification |
| 本文献已被 维普 等数据库收录！ |
