基于不确定信息系统的动态属性约简算法

Dynamic Attribute Reduction Algorithm Based on Uncertain Information System

针对不确定信息系统中存在的问题,在不确定信息系统的属性集上定义了一种等价关系,以等价类决定属性的条件信息量,从而定义属性的相对重要性,给出一种新型的动态属性约简算法。在一个信息系统不断变化时,该算法不像静态约简算法需要重新计算,而是利用新增的数据对已有的数据进行修正,使得每次约简充分利用上一次的结果,这样使得同样的比较运算不会重复进行。通过分析表明,该算法是正确有效的。

For uncertain information system about reduction problems, we proposed a dynamic attribute reduction algorithm,which in the uncertain information system of the set of attributes define an equivalence relation, with equivalent sort of decision the conditions information of the attribute and take the conditions information define the relative importance of attributes. When the number of the object in the information table increases, instead of treating the changed information table as a new one and need to recount, just updates the old information based on the increased objects, this makes the same comparison operations not repeated. By the example analysis shows that the algorithm is correct and effective.