提出了一种新的连续信息系统的知识获取方法.首先给出了连续属性离散化方法,该方法可以使得离散化后得到的分类的类数适度,丢失的信息量尽量减少;然后在此基础上给出了连续信息系统属性约简的一种简易算法;最后实例证明了该离散化方法是可行和有效的.
This paper proposes a new knowledge acquisition approach in continuous information systems. Firstly, a new method of discretizing continuous attributes is put forward, which can make the number of the obtained classes more moderate and the lost information fewer. And then a simple attribute reduction algorithm is developed in continuous information systems. At last, an example is used to illustrate that the approach is feasible and effective.