一种动态时间阈值的会话识别算法

A Conversation Recognition Algorithm of Dynamic Time Threshold

会话识别是网络日志预处理中的重要环节,传统网络日志数据的预处理采用固定阈值会话识别算法,不能较好消除网络日志中的冗余信息。为更好提高后继数据挖掘的效率和精度,提出一种改进的预处理网络日志数据的会话识别算法,根据内容及站点结构确定其重要程度,对时间阈值进行动态控制调整,根据用户对访问内容的兴趣度删除不感兴趣的页面,消除冗余信息。

Conversation recognition is an important link in the pretreatment web log. The traditional network log datapretreatment use the fixed time threshold conversation identification algorithms, it can＇t eliminate the redundant information web logs. We put forward an improved network log data pretreatment of conversation identification algorithms in order to improve the efficiency of data mining and subsequent accuracy. According to the content and structure to determine its important degree site. Improving the time threshold in a dynamic control adjustment. According to the user＇ interests to eliminate the redundant information.