一种正区域的递归计算方法

A Recursive Method of Computing Positive Regions

正区域的计算在粗糙集的应用中十分重要,传统的求解方法较为复杂,计算量大。本文在深入研究粗糙集基本理论的基础上,得出了一种等价的计算正区域算法。该算法对论域中的样本进行分步分类,同时判断其是否属于正区域,而后将其从论域中删除,从而减轻计算量,并同时求出不属于正区域的样本集。然后分析了新增加的属性不会对已有正区域集合运算产生影响,因此只需要对原来的不属于正区域的样本集进行进一步分析就可以得到增加属性后的正区域,从而使计算正区域的范围逐渐减小,得到了一种递归计算正区域的方法。理论分析和实例结果验证了该计算方法的有效性和可行性。

The computing of positive regions is very important in the application of rough sets. Traditional computing methods are more complex. So based on all in-depth study of the rough set theory, a equivalent method for computing positive regions is proposed. The algorithm classifies the samples in universe step by step and determines whether those samples belong to the positive region and deletes the samples from the universe to reduce computational complexity and gets the un- positive region set. Then, a method of recursively computing the positive region is proposed by analyzing the influence of the increased attributes to the set of known positive regions. ,So the positive region of an increased attribute system can be gotten by analyzing the un-positive region, And the theoretical analysis and experimental results show that this new method is much more efficient and useful in computing.