KDD中知识的自动评价系统

An Automatic System for Evaluating the Discovered Knowledge in KDD

1.引言 在知识发现过程中,通过挖掘算法产生大量的模式,例如斯坦福大学对一个有3万条记录的调查数据库中使用关联规则的发现算法,产生了2万多条的规则[1],而大多数规则是显而易见和毫无意义的.

How to automatically evaluate the discovered knowledge is very important in KDD,but the research on this problem is very little- The paper gives one automatic system for evaluating the knowledge, and provides many solutions. First some relative concepts are described,and the construction of this system is given,and uses the case to prove it.