面向属性归纳的特征规则挖掘方法的改进

Improvement of mining method with attribute-oriented induction characteristic rules

在浩瀚的数据资源中发现并提取有价值的知识是目前数据库、人工智能等学科研究中的共同热点问题.面向属性的归纳是一种广泛使用的特征规则的发掘算法.本文对特征规则挖掘的面向属性归纳方法进行改进.提出了"分类信息库"的概念.给出了基于这种概念的新算法.利用这种算法可以发现很多原方法发现不了的新知识,使特征规则的发现有了更大的实用价值.实例分析表明,该算法能够准确地发现特征规则.

Altribute-oriented inducing algorithm is improved.The concept of 'classing information database' is introduced,based on which a new method is proposed.Much new knowledge could be discovered with the new method.It demonstrates that mining characteristic rules is of more practical value.Using the new method we can find much of new knowledge that can`t be fond before.So it becomes more practical value to find the mining characteristic rule.