基于IP的数据挖掘初探

现有IP的查询、统计和报表，其处理方式大都是对指定的数据进行简单的数字处理，而不能对这些数据所包含的内在信息进行提取。要想挖掘深层次的信息就需要数据挖掘技术。研究了如何增加用户上网时长的问题。提出了研究的框架和步骤，首先收集数据，接着把和用户上网时长相关的因子提取出来，通过多种数据统计分析方法，发现这些因子之间的内在相关性，并通过一定渠道检验模型的正确性。

Primary Investigation Data Mining on IP

The query, statistics and report about IP are most simple digital processes to special data in recently. These processes can not obtain knowledge from data. Data mining is a method to get model from lots of data. This paper points out the mainframe and process about how to increase user's hours on the net. First, we pick up correlative data. Secondly,we distill correlative data about user's hours on the net. Thirdly, we obtain principle through many statistical analysis methods. Finally, the correctness is tested through some methods.